Skip to main content

Full text of "The Astrophysical journal"

This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project 
to make the world's books discoverable online. 

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject 
to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books 
are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover. 

Marks, notations and other marginalia present in the original volume will appear in this file - a reminder of this book's long journey from the 
publisher to a library and finally to you. 

Usage guidelines 

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the 
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we have taken steps to 
prevent abuse by commercial parties, including placing technical restrictions on automated querying. 

We also ask that you: 

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for 
personal, non-commercial purposes. 

+ Refrain from automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine 
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the 
use of public domain materials for these purposes and may be able to help. 

+ Maintain attribution The Google "watermark" you see on each file is essential for informing people about this project and helping them find 
additional materials through Google Book Search. Please do not remove it. 

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just 
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other 
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of 
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner 
anywhere in the world. Copyright infringement liability can be quite severe. 

About Google Book Search 

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers 
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web 



at |http : //books . google . com/ 



LIBRARY 

UNIVis-RSi i Y OF CALIFOimTA 
DAVIS 




THE 

ASTROPHYSICAL JOURNAL 

An International Review of Spectroscopy and 
Astronomical Physics 



Volume I 
189s 



Reprinted with the permission of the original publishers 

JOHNSON REPRINT CORPORATION 
NEW YORK, NEW YORK 



FRONTISPIECE 




JiotjtM 



PHOTOGRAPH NEAR MESSIER ii 

i8g2t June ag^ tt^ 40^-15^ j*" Pacific Standard Time 

Taken by E. E. Barnard with the 6-inGh WilUrd Lens of the Lick Observatory 



THE 

AsTROPHYSICAL JoURNAL 

An International Review of Spectroscopy and 
Astronomical Physics 

EDITORS ) 

GEORGE E. HALE JAMES E. KEELER 

Dirtcior of the Yerkes Observatory Director of the Allegheny Ohervatory 

ASSISTANT EDITORS 
J. S. AMES HENRY CREW 

Johns Hopkins University Northwestern University 

W. W. CAMPBELL E. B. FROST 

Luk Observatory Dartmouth CoUege 

F. L. O. WADSWORTH, University of Chicago 

ASSOCIATE EDITORS 

M. A. CORNU C. S. HASTINGS 

i&coie Fotytechnique, Paris Yale University 

N. C. DUNl&R A. A. MICHELSON 

Asironomiska Observaiorium, Upsala University of Chicago 

WILLIAM HUGGINS E. C. PICKERING 

Tube Hill Observatory^ London Harvard College Observatory 

P. TACCHINI H. A. ROWLAND 

Osservatorio del Collegio Romano, Rome Johns Hopkins University 

H. C. VOGEL C. A. YOUNG 

Astrophysikalisches Observaiorium, Potsdam Princeton University 



VOLUME I 
JANUARY— MAY 1895 



CHICAGO 
1895 



LIBRARY 

UNIVERSITY OF CALIFORNIA 
DAVfS 



First ripriiUiing, iQSQt Johnsttn Refrini Corporahon 



TABLE OF CONTENTS. 



GENERAL. 



A Cloud-like Spot on the Terminator of Mars. A. E. 

Douglass 127 

A Combination Telescope and Dome. A. £. Douglass • - 401 
A Spectroscopic Proof of the Meteoric Constitution of 

Saturn's Rings. James £. Keeler 416 

Comparison of Photometric Magnitudes of the Stars. £. C. 

Pickering 154 

Discovery of Variable Stars from their Photographic 

Spectra. E. C. Pickering 27 

Eclipse OF Jupiter's Fourth Satellite, February 19, 1895. £. 

C. Pickering 309 

Note on the Atmospheric Bands in the Spectrum of Mars. 

William Huggins 193 

Note on the Spectrum of Argon. H. F. Newall - - - 372 
Observations of Mars Made in May and June, 1894, with the 

Melbourne Great Telescope. R. L. J. EUeiy • - - 47 
On a Lens for Adapting a Visually Corrected Refracting 

Telescope to Photographic Observations with the 

Spectroscope. James E. Keeler loi 

On a New Form of Spectroscope. C. Pulfrich - - - . 335 
On a New Method of Mapping the Solar Corona without an 

Eclipse. Geoige E. Hale 318 

On a Very Large Protuberance Observed December 24, 1894. 

J. F^nyi 212 

On Martian Longitudes. Percival Lowell 393 

On Some Attempts to Photograph the Solar Corona without 

AN Eclipse, Made at the Mount Etna Observatory. A. 

Riccd 18 

On the Conditions which Affect the Spectro- Photography of 

the Sun. A. A. Michelson i 

On the Distribution of the Stars and the Distance of the 

Milky Way in Aquila and Cygnus. C. Easton - - - 216 
On the Periodic Changes of the Variable Star Z Herculis. 

N. C. Dun^r 285 

iii 



IV CONTENTS 

rAGS. 

On the Spectrographic Performance of the Thirty-inch 

PuLKOWA Refractor. A. B^lopolsky 366 

Photographic Observations of Eclipses of Jupiter's Satel- 
lites. Willard P. Gerrish 146 

Photographs of the Milky Way. £. £. Barnard - - - 10 

Preliminary Table of Solar Spectrum Wave-lengths. I, p. 29 ; 
II, p. 131; III, p. 222; IV, p. 295; V, p. 377. Henry A. 
Rowland 

Recent Changes in the Spectrum of Nova AuRiGiS. W. W. 

Campbell 49 

Recent Researches on the Spectra of the Planets. I, p. 196 ; 
II, p. 273. H. C. Vogcl 

Remarks on Professor £. C. Pickering's Article, "Compari- 
son OF Photometric Magnitudes of the Stars," in A. N, 
3269. G. MQIIer and P. Kempf 428 

Schmidt's Theory OF THE Sun. £. J. Wilczynski - - -112 

Solar Observations Made at the Royal Observatory of the 

Roman College in 1894. P. Tacchini .... 210 

Stars Having Peculiar Spectra. £leven New Variable 

Stars. M. Fleming 411 

T ANDROMEDiB. £. C. Pickering 305 

The Arc-Spectra op the Elements. I. Boron and Beryllium. 

H. A. Rowland and R. R. Tatnall 14 

II. Germanium 149 

The Modern Spectroscope. X. General Considerations 
Respecting the Design of Astronomical Spectroscopes. 
F. L. O. Wadsworth 52 

The Modern Spectroscope. XI. Some New Designs of Com- 
bined Grating and Prismatic Spectroscopes op the 
Fixed-Arm Type, and a New Form of Objective Prism. 
F. L. O. Wadsworth - - 232 

The Modern Spectroscope. XII. The Tulse Hill Ultra- 
violet Spectroscope 359 

The Spectrum of ^ Cephei. A. B^lopolsky 160 

The Spectrum OP Mars. Lewis £. Jewell 311 



MINOR CONTRIBUTIONS AND NOTES. 

Arthur Cowper Ranyard. George £. Hale - - - - 168 

A Large Eruptive Prominence. George £. Hale ... 433 
A Large Reflector for the Lick Observatory. Edward S. 

Holden 442 



CONTENTS V 

rAGB. 

Device for Putting Wave-lengths on Spectrum Plates. Olin 

H. Basquin i66 

Dr. Pulfrich*s Modification of the Littrow Spectroscope. 

James E. Keeler 353 

Meeting of the Section of Mathematics, Astronomy and 

Physics of the Chicago Academy of Sciences, December 

11,1894. T.J. J. Sec 86 

Note on the Arc-Spectrum of Copper. H. Kayser - - - 84 
Note on the Exposure Required in Photographing the Solar 

Corona without an Eclipse. George E. Hale - - - 438 
Notes on Silvering Solutions and Silvering. F. L. O. Wads- 
worth - 252 

On Brester's Views as to the Tranquillity of the Solar 

Atmosphere. Egon von Oppolzer 260 

On a Photographic Method of Determining the Visibility of 

Interference Fringes in Spectroscopic Measurements. 

George E. Hale 435 

On Determining the Extent of a Planet's Atmosphere W. W. 

Campbell 85 

On the Variability of Es.-BiRM. 281. T. E. Espin - - - 351 
Photographic Correcting Lens for Visual Telescopes. James 

E. Keeler 350 

Schmidt's Theory of the Sun. James E. Keeler - - - 178 

Spectro-Bolographic Investigations at the Smithsonian 

Astrophysical Observatory. George E. Hale - - - 162 
The Astrophysical Journal. George E. Hale .... 80 
The Color of Sirius in Ancient Times. W. T. Lynn - - -351 
The Design of Astronomical Spectroscopes. James E. Keeler 248 
The Design of Electric Motors for Constant Speed. F. L. 

O. Wadsworth 169 

The Displacement of Spectral Lines Caused by the Rotation 

OF A Planet. James E. Keeler ...-.- 352 
The Short Wave-lengths of the Spark Spectrum of Aluminium. 

C. Runge 433 

The Variable Star 3416 S Velorum. James E. Keeler - - 262 

Terrestrial Helium (?) 439 

Wolsingham Observatory Circular No. 41. T. E. Espin - - 87 



REVIEWS. 



Beitrage zur Kenntniss der Linien-Spectren ; J. R. Rydberg, 

by J. S. Ames 90 



vi CON7ENTS 

PACK. 

Beitrage zur Kenntniss dbr Linien-Spbctren ; H. Kayser u. C. 

Runge, by J. S. Ames 90 

£tudb sur lb spectre db l'6toile variable a Cephbi; a. 

B61opolsky, by E. B. Frost 263 

Flame Spectra at High Temperatures. II and III ; W. N. 

Hartley, by J. S. Ames 89 

On the Spectrum of the Electric Discharge in Liquid Oxygen, 

Air and Nitrogen ; Liveing and Dewar, by J. S. Ames - 88 

On Variations Observed in the Spectra of Carbon Elec- 
trodes, AND ON the Influence of One Substance on the 
Spectrum of Another ; W. N. Hartley, by J. S. Ames - 88 

Popular Scientific Lectures | Ernst Mach, by Henry Crew- - 267 

Preliminary Report on the Results Obtained with the Pris- 
matic Camera During the Total Eclipse of the Sun, 
April 16, 1893 ; J. Norman Lockyer (abstract), by £. B. Frost 91 

Publications of the Lick Observatory, Volume III, by George 

E. Hale 180 

The Luminosity of Gases, III ; A. Smtthells, by Henry Crew - 266 

The Source and Mode of Solar Energy Throughout the Uni- 
verse ; I. W. Heysinger, by James E. Keeler ... 368 

The Spectrum Researches of Professor J. M. Eder and E. 

Valenta ; by J. S. Ames 443 

The Total Solar Eclipse of April 16, 1893. Report on 
Results Obtained with the Slit Spectroscopes ; E. H. 
Hills, by E. B. Frost 91 

Uebbr die Spectra von Zinn, Blei, Arsen, Antimon, Wismuth ; 

H. Kayser u. C. Runge, by J. S. Ames 91 



RECENT PUBLICATIONS .... 93, 189, 270. 354» 447 



THE 

ASTROPHYSICAL JOURNAL 

AN INTERNATIONAL REVIEW OF SPECTROSCOPY 
AND ASTRONOMICAL PHYSICS 



VOLUME I JANUARY 189J NUMBER 1 



ON THE CONDITIONS WHICH AFFECT THE 
SPECTRO-PHOTOGRAPHY OF THE SUN. 

By Albert A. Michelson. 

The recent developments in solar spectro-photog^phy are 
in great measure due to the device originally suggested by Jans- 
sen and perfected by Hale and Deslandres, by means of which 
a photog^ph of the Sun's prominences may be obtained at any 
time as readily as it is during an eclipse. The essential features 
of this device are the simultaneous movements of the colli- 
mator-slit across the Sun's image, with that of a second slit (at 
the focus of the photographic lens) over a photographic plate. 
If these relative motions are so adjusted that the same spectral 
line always falls on the second slit, then a photographic image 
of the Sun will be reproduced by light of this particular wave- 
length. 

Evidently the process is not limited to the photography of 
the prominences, but extends to all other peculiarities of struct- 
ure which emit radiations of approximately constant wave- 
length ; and the efficiency of the method depends very largely 
upon the contrast which can be obtained by the greater enfeeble- 



2 ALBERT A, MICHELSON 

ment of the background of white light, by dispersion, as com- 
pared with the relatively fixed intensity of the homogeneous 
radiations. 

An aid to the practical realization of the method, whose 
importance can hardly be overestimated, consists in the use of 
the bright H and K lines for this purpose ; their exceptional 
efficiency depending upon the fact that the continuous spectrum 
in their immediate vicinity is already very much reduced by 
broad absorption bands. 

In addition to this contrast, it is of course necessary that the 
actual intensity of the light should be sufficient to affect the 
photographic plate, and also that the outline and details should 
be as sharply defined as the photographic process and other 
conditions permit. These conditions and their effects upon 
these essential circumstances we now proceed to investigate. 
Let s = width of collimator slit, 

h = length of the image of a prominence, 
X, and A„ their spectral images, 

)8 = -^ = aperture of objective, 

ft = ^ = aperture of collimator, 

o = — ' = aperture of photographic lens, 

/• 
^, = angle of incidence, 
^, = angle of diffraction, 

<r = — = gratmg space, 

m = order of spectrum. 
If the whole aperture of the collimator is filled With light by 
the cone of rays from the objective, then ft = fi. The magnifi- 
cation of the spectroscope is^; and there is a further enlarge- 

ment due to diffraction by the aperture of the photographic 
lens (to which should be added the effects of imperfections in 
the lenses and gratings or prisms, of vibrations and of inequali- 
ties of temperature); and, finally, the enlargement by dispersion. 



SPECTRO-PHOTOGRAPHY OF THE SUN 3 

Supposing the effective aperture to be square,* the effect of 
diffraction, the source being a point, is given by the formula 

sin' ^ X 

1=——-, in which ♦=--ojc, jr being the distance from the 

geometrical image. 

This is shown graphically in the figure, in which the ordi- 
nates represent intensity and the abscissae the angle ^. 

The actual width of the line is that for which ^ = 2x ; that is, 

jc,= 2- ; but the apparent width will vary with the sensitiveness 
of the eye, or the photographic plate. 




In the case of a source of finite width, the effect would be 
found by integrating this expression between values correspond- 
ing to the width of the source. In any case the geometrical 

image is broadened by a quantity e-, in which (in view of the 

uncertainty just noted, and considering also that the various 
disturbing elements previously noted make the actual value still 
more vague) c may represent a factor which lies between 1.5 
and 2.0. 

We have next to investigate the dispersion, and for this pur- 
pose it will be convenient to assume that it is produced by a 
grating. Suppose the light of the source to have a range of 

wave-length 5X. The dispersion will then be ^8A. But 

d^ __ m 
^A^o-cos^, ' 

' The result for a circular aperture would be to increase the width one-fourth. 



ALBERT A. MICHELSON 



and substituting for <r its value, we have for the enlargement due 
to dispersion mnl\ 



^.cos^. ' 

Remembering that the effective aperture of the photographic 
lens is d^ cos <^., d^ being the length of the grating which is 

covered by the incident light ; and that ^,= ^, we obtain 

for the width of the spectral image of a slit of width s the 
expression --^ a. / ^cos<^. , . mnl\ 

^•"//■^ *^- ^sCos<^. +^- ^.cos.^. • 
For our present purpose it will not be necessary to include 
the effect of the angles ^^ and <^„ which we may suppose so 
small that their cosines are nearly equal to unity. 

Substituting for -^ its approximate value ft and for^its.ap- 

proximate value - we have 

J. = -()3j+ eX+ ««8X). (i) 

If the intensity of the source (diminished by absorption 
through the atmosphere) be represented by unity, and e repre- 
sent a factor depending on losses in the spectroscope other than 
those due to diffraction, dispersion and magnification, then we 
have for the intensity of the spectral image 

e , ^P sh 
I = - sin* - — 7- , 

2 2 J, A. 

or very nearly i = g (ff A + cA)y+lx+ ^^8A) ' (^> 

If we may extend the assumption that the intensity of a 
spectral "line" is inversely proportional to the ratio of its breadth 
to the entire effective portion of the spectrum, then supposing the 
background to be / times as bright as the approximately mono- 
chromatic source, we have for the ratio of intensities the expres- 
sion _ ^_ )5j+eA+««8A 






SPECTRO'PHOTOGRAPHY OF THE SUN 5 

X, and A, being the limits of the effective portion of the spectrum. 
From this it appears that it will be of advantage to increase 
the dispersion (m fi) until 

fis'\-€\^\mnh\. (4) 

It is therefore important that the first slit should be narrow ; 

but no great advantage will be gained when -f < i — , while 
beyond this point light is lost by the diffracted rays falling out- 
side the collimator. « . 
With the above value of ninl\ we have r = -/t r-. 

2 A, — A, 

By doubling this dispersion the ratio r would only be reduced 
in the proportion of 5 : 6, while the loss in sharpness and bright- 
ness would much more than compensate for this slight increase 
in contrast. 

These effects we now proceed to investigate. 

The sharpness of the photographic image (supposing the 
focusing to be exact and the effects of irradiation negligible) 
will depend on the distinctness of the spectral image which falls 
on the second slit. This slit should be narrow (simply to pro- 
tect the parts of the plate where there is no image, but merely 
the spectrum of the white background), but no important advan- 
tage will be gained by making the width less than that of the 
image of the first slit. 

The sharpness of detail in the spectral image is independent 
of the slit-width, except in so far as this entails increased dis- 
persion in order to maintain contrast. 

If the source were a line, its first image would have a width' 

-T-T. This image is magnified by the spectroscope in the pro- 

portion ~» and is further enlarged' by the quantity - (eX+ « «8X); 
so that the actual hazy border may be measured by 

' DUreguding the accidental sources of indistinctnef s already mentioned. 

"The expression for the increase in width due to dispersion assumes that within 
the limits X to X + ^X the intensity* is constant. 

In view of the uncertainties mentioned above, together with the difficulty in ascer- 
taining the true distribution of intensity as a function of X, it seems scarcely worth while 
to perform the double integration which would lead to the more accurate expression. 



ALBERT A. MICHELSON 
wz=z^{2€k + mn&\y is) 



a 



If the slit have the minimum efficient width (s= — —) 

^ 2 ft ' 

then the most suitable dispersion is wr8X = 3 cX, i^hich gives 

w, = ; so that the effect of dispersion in this case is to 

increase the border-width 2j4 times. This increased width is, 
however, only about o'""'.038, which in a 50 to 100"" disk may be 
safely ignored. 

If, however, the slit be as wide as o"".4, the dispersion 
m»8A would have to be increased to o™".04 15 to maintain the 
contrast, and w would then be over o"'".4o, an amount of 
indistinctness which would make a poor image. 

For the absolute brightness (supposing the prominence to 
cover the whole width of the slit) we have 



t = - 



which shows that the effect of the given dispersion is to diminish 
the intensity to one-third of the value it would have with zero 
dispersion. This is not, however, so serious a loss as to make 
it worth while to sacrifice any of the contrast ; but this would 
hardly be true with four times as great a dispersion, for in that 
case the intensity would have only one-ninth its maximum value.' 
The conditions for the limit of useful dispersion are given by 
the formula 

If the first slit be so narrow that fis may be neglected beside 

e X (for instance, if J < o""".oo5), then « « = -^ = ^ in tenth - 

meters. 

Thus, if 8X = o.25 tenth-meters (which is about the "width" 

of the reversed calcium lines), mn = 60,000; so that, if the grat- 

' The formula shows how important it is that a, the aperture of the photographic 
lens, should be as large as possible. Without any losses by reflection, absorption or 
dispersion, the maximum intensity of the image is tiH = ^a', and since it is practically 
impossible to make a ]> i, we have im^-whi' '^^^ inevitable losses in any instrument 
with high dispersion would reduce this value, ten times ; so that with the effect of 
dispersion it is not likely that f should ever exceed fiiW* 



SPECTRO-PHOTOGRAPHY OF THE SUN 






W 

I o 

IS 



1 



If 

SI 



o M M m iN.ao 



;§§§! 



XoMfomtooo 
ei M M M ^ r«. d 



Jo o M ror^QO Oi 


^§§§§§1g 






M M O M ei N M 



88 



a too 00 o o to m 
• O O O •-» fnmo 



< orOroQOO^r^ 
•;• o 1^ ^«o QO o» c> 

i 



^§§§§§§1 



ft t^ r«. M N M M 
_p CO *0 *030 00 <*> CO 

• oooocitno 



. So^SS,gg 

o d d d d M ci 



ing have 6o,ooo lines, the first spectrum 
would give sufficient dispersion. 

If, however, the slit be so wide that cX 

may be neglected, then « « = ^-^, or, 

o I O.I J 

supposmg ^ =-.«« = -j^ . 

Thus suppose ^ = 0"".!, then 
mH = 4X 10*. 
That is, a grating ot 100,000 lines, in the 
fourth spectrum, would just suffice to give 
the necessary dispersion. For a still larger 
slit, it would be practically impossible to 
use too high a dispersion ; but as shown 
above the indistinctness of the image 
under these circumstances would border 
on the limits of toleration. 

If in the preceding formulx we substi- 
tute , , 

X, — A,= 2500 tenth-meters, 

8A=:o.25 tenth-meters, 

fi=z .05, a= .10, e= 1.5, A = o"".ooo5, 

we get for the width of the spectral line 

for the ratio 

r_ 200 J -f 3 
/"" mn 



-j-.oooi 



for the intensity 

800 - = ; 

e J-J-.015 + 5 X io"^« « 

and for the width of hazy border 
a/ =.015-1- ^'5 X lo'^mn. 

From these formulae the accompanying 
table was computed. 

If the dispersion is produced by a 
train of prisms instead of a grating, we 

have only to substitute for -j- in these 



8 ALBERT A. MICHELSON 

expressions, the equivalent value for the dispersion of a prism ; 

du, t 
namely, ^3-. in which fi is the index of refraction, and / the 

difference in thickness of glass traversed by the extreme rays. 

If the prisms are of 60° then / is roughly 2pd^,p being the 

du, 
number of prisms; and in this case the dispersion is ^/^ • 

d u, 
For the ordinary varieties of flint glass ^ is about 2 X lo"*, 

if ^X is expressed in tenth-meters. If the grating space- (sup- 
pose o"".oo25) be expressed in the same units, then for equal 
dispersion 4X io-«/ = 4X id^m; or the number of prisms *is 
equal to the order of the spectrum. 

It will be of interest to compare the intensities of the grat- 
ing spectrum with that of the equivalent train of prisms. If 
the grating acts by opacity alone (see Lord Rayleigh's article 
on Wave Theory of Light in Enc, Brit,), then the maximum 

brightness of the mth spectrum is — ^ of the incident light. On 

the other hand if p represents the fraction of incident light 
which is transmitted by a single prism, then the intensity of the 
light which has passed through / prisms is p^;' and if these 
are to be the same wc have p^ = «r^ w^ • 

The fraction p varies considerably with the material and the 
construction of the prism. It will be somewhere near the aver- 
age (though probably under rather than over), if it is placed at 
0.5 ; in which case we have approximately 2"*= 10 «•. 

Suppose the grating to contain 400 lines to the mm. Then as 
just found, /=m; and the equation is nearly satisfied if /s=io. 
Accordingly, as many as ten prisms might be employed before 
the superiority of the higher orders of spectra would be manifest 
If the grating has 800 lines to the mm, p=:2m, and the advantage 
would be with the prisms up to/ =7; while for «, = 1200 lines to 
the mm (which has been found by experience to be near the 
practical limit of efliciency),/ = 3M and the equation is most 
nearly satisfied by / = 3. 

' Professor Pickering has pointed out that this loss is considerably diminished in 
consequence of the partial polarization of the transmitted light. 



SPECTRO-PHOTOGRAPHY OF THE SUN 9 

It will be noted that in every case the prisms have the advan- 
tage up to the highest observable spectrum of the grating. This 
depends of course on the value assumed for p, which is in many 
cases undoubtedly too low. On the other hand, it is well known 
that the law given for the intensity of the grating spectra holds 
only when it acts by opacity ; while in fact gratings are now 
made which throw a much larger proportion of light into some 
one spectrum. 

It appears on the whole that there is little ground for a choice. 
It may be noted, however, that while in the case of the grating 
the theoretical limit is nearly reached, the prism is capable of 
still further improvement in the direction of material having 
great transparency and homogeneity, together with high disper- 
sion. 

In the practical application of the principles here discussed, 
many difficulties may arise owing to peculiarities of photographic 
processes. Thus if the plates, or the process of developing, 
employed be very sensitive to small differences in intensity when 
the light is bright, some of the "contrast" may be sacrificed. 
On the other hand, if the plates are nearly equally sensitive 
through a considerable range of small intensities, then the object 
should be to gain in contrast by increasing the dispersion, even 
at the cost of brightness — until the sharpness of the image 
begins to deteriorate. 

Here again it is possible that the structure of the film may 
define the limiting condition, either by the coarseness of its 
grains, or by the amount of spreading of the photographic effect 
by irradiation. 



PHOTOGRAPHS OF THE MILKY WAY. 

By E. E. Barnard. 

In my photographic survey of the Milky Way with the 6-inch 
Willard lens of this Observatory, I have come across many very 
remarkable regions. Some of these, besides being remarkable 
for showing the peculiar structure of the Milky Way, are singu- 
larly beautiful as simple pictures of the stars. I have selected 
two of these for illustration in The Astrophysical Journal. 

THE REGION OF MESSIER II. 
(Frontispiece.) 

Every telescopic observer is familiar with the beautiful cluster 
known as Messier Eleven (R. A. iS** 45"; Dec. S. 6** 24'), for it 
is an easy object with almost any instrument. With a low power 
on a considerable telescope M. 1 1 is one of the prettiest clusters 
in the sky. It is a rather open non-condensed cluster of 11-12 
magnitude stars grouped about a 9 magnitude star, and covering 
a space of about 10' or 12'. Like the larger and denser clusters, 
this also has vacant spaces in it. 

From its great beauty with a low power on the 12-inch, I have 
often shown it to visitors here on Saturday nights. From an 
optical illusion, they have invariably seen "millions of stars in 
it." 

However, there is not a vast number of stars actually com- 
posing this cluster, and it would not be a difficult task to count 
them. One remarkable thing in connection with the expressions 
of the visitors when looking at M. 1 1 is that a considerable per- 
centage of them instinctively call attention to the form of the 
cluster itself as being that of a star. In my experience thus, I 
think there will have been from fifty to bne hundred people who 
have independently exclaimed at its stellar form. 

In the telescope, the cluster itself is seen to be projected on a 
uniform background of very small stars, and anyone examining 



PHOTOGRAPHS OF THE MILKY WAY II 

the region about it would not be specially impressed with the 
neighborhood. But looking along the telescope tube one sees 
that it is placed in one of the great luminous clouds of the Milky 
Way. 

In my work on the Milky Way this cloud was one of the first 
objects photographed, and I have several pictures of it, the finest 
of which I herewith present to the readers of The Astrophysical 
Journal. 

Let us examine this picture carefully. The plate has taken 
in the entire cloud that had been seen with the naked eye. The 
scale of the picture is small, consequently M. 1 1 itself is simply 
a spot of light — the individual stars not being seen separately. 
It is seen to 'occupy the upper north edge of the great naked- 
eye cloud, and appears like a nucleus to that object. I think it 
hardly questionable but that M. 1 1 is really a nucleus to the greater 
cluster, though it may be only a simple case of projection. 

The naked-eye cloud itself, however, becomes on the photo- 
graphic plate a vast and gorgeous cluster of stars. It is shown 
to be an immense irregular cluster of apparently very small stars, 
and is seemingly perfectly isolated from the rest of the Milky 
Way. 

In looking at this picture several people have called attention 
to the fact that when held in certain positions, the outline of 
this star cloud has a decided resemblance to some pictures of the 
great nebula of Orion. To the west of M. ii the star cloud is 
brightest and most definite in form, with several detached out- 
lying portions. 

Running southerly from M. ii is a broad curving semi-vacancy. 
About 3° south and west of M. ii — from a large semi-vacant 
region — two similar partial vacancies run divergingly eastward, 
and, joining the first mentioned dark stream from M. ii, form 
a distinct A. The northern of these thin streams partially cuts 
off the brightest mass of the cloud to the west of M. ii. 

In looking at this picture, I have often received the impres- 
sion that this huge cloud of stars had been generated by some 
tremendous whirling motion. 



12 E. E. BARNARD 

There are other curious and interesting features shown on 
this picture, such as vacancies and lines of stars; though I 
must confess that there is a marked absence here of the geomet- 
rical lines and curves that form such a striking feature in other 
portions of the Milky Way. Indeed, it seems to be a fact that 
this geometrical arrangement of the stars is more or less absent 
in all the definite clouds of stars, and is remarkably conspicuous 
in the larger uniform areas. 

THE MILKV WAY NEAR CHI CYGNI. 
(Plate II.) 

I have before, elsewhere, called attention to the fact that the 
Milky Way does not anywhere repeat itself. That is, there are 
no two parts of the Milky Way that seem to be made up on the 
same plan. 

Certain regions are formed by coarse stars mainly. Such is the 
region east and north of Orion, where these coarse stars are 
nearly uniformly scattered. Again, there are the great cloud 
regions where the massing clouds are made up uniformly of 
extremely small stars — that is the stars are apparently, and I 
believe are really very small. Such for instance are the great 
clouds in Sagittarius in R. A. i8^; Dec. S. 28"". 

This region near Chi Cygni seems to be made up of a back- 
ground of fairly small stars with a decided sprinkling of larger 
and coarser ones. 

The picture divides itself diagonally into two regions. The 
northwest half is a dense matting of small stars too thick and too 
dense to be seen through — forming thus a perfect veil over that 
part of the sky. This suddenly shoals diagonally along the middle 
of the plate and leaves the southeast portion covered uniformly 
with a thin sheeting of small stars, through which we readily 
look out into the blackness of space beyond. In this thin region 
one easily picks out the geometrical tracery, which is seemingly 
made up of a vast number of vacant paths among the stars. 

A little to the west of the center a great semi-vacancy is seen 
in the dense region of stars ; this has a parabolic form and runs 



PLATE II 




PHOTOGRAPH NEAR CHI CYGNI 

iSgit October *o, 6* 47*^ -ti^ /?"• Paci/ic Standard Titue 

Taken by E. £. Barnard with the 6-inch Willard Lens of the Lick Observatory 



PHOTOGRAPHS OF THE MILKY WAY 1 3 

southwesterly. About one degree above this and slightly to the 
west is a curious small and sharply defined A shaped semi- 
vacancy. The star Chi occupies a position about the middle of 
the plate, and from its reddish color is lost among the small 
stars, though it is easily found with a diagram. The bright star 
in the northeast corner of the plate is y Cygni. This star is sur- 
rounded by great masses of nebulosity on the original plate ; 
but to bring out the peculiar structure of the Milky Way itself in 
this region it was necessary to neglect the nebulosity, which to 
be well shown would require another positive from the same 
negative. 

In examining this plate one will see, also, that the geomet- 
rical patterns of this part of the sky are also carried over on to 
the dense matting of stars occupying the westerly part of the 
plate, though they are more marked to the east, where the stars 
are thinly distributed. 

Looking at these two pictures, one who is familiar with these 
two regions telescopically, cannot help but marvel at the won- 
derful power of the photographic plate over that of the eye and 
the telescope alone, in dealing with that magnificent zone of 
stars — the Milky Way. 

Mt. Hamilton, Dec. 12, 1894. 



THE ARC-SPECTRA OF THE ELEMENTS. I. 

BORON AND BERYLLIUM. 

By Henry A. Rowland and Robert R. Tatnall. 

This paper is intended as a preliminary notice of a series of 
investigations on the arc-spectra of certain of the elements which 
have not hitherto been carefully studied by modern methods. The 
investigations may be regarded as a continuation of the work 
upon the solar spectrum commenced several years ago by one of 
us, inasmuch as the ultimate object in view is the identification 
of some of the many lines in the spectrum of the Sun, whose 
origin still remains unknown. 

That portion of the work here described will be confined to 
the visible and ultra-violet portions of the spectrum. The infra- 
red portion forms the subject of another investigation which is 
now being carried on in this laboratory, the results of which 
are expected to appear later on. 

The plates to be used will be for the most part selected from 
the series made some years since in connection with the study 
of the solar spectrum.* They are mostly 19 inches in length, 
and were made with a six-inch concave grating of 21^ feet 
radius, ruled with 20,000 lines to the inch. They comprise 
spectra of nearly all known elements, in both first and second 
orders, the latter being accompanied with the solar spectrum for 
purposes of comparison, according to the well-known method of 
Rowland.' 

In studying the spectrum of any element, it is first necessary 
to eliminate all lines due to impurities by a careful comparison 

' The completeness of this set of plates is partly due to financial aid received from 
the Bache Fund of the National Academy of Sciences, from the fund given by Miss 
Bruce to the Harvard Astronomical Observatory for the prosecution of astrophysical 
research, and from the Rumford Fund of the American Academy of Arts and 
Sciences. 

■See Ames: "The Concave Grating in Theory and Practice," A, and A, XX, 28, 
1892. 

14 



ARC-SPECTRA OF BORON AND BERYLLIUM 1 5 

with the spectra of the carbon poles, and of all elements likely 
to be associated with the substance under examination. This 
comparison is made by direct superposition of the plates, which 
are all to nearly the same scale, thus insuring almost perfect 
coincidence of corresponding lines. 

The measuring engine employed has a nearly perfect screw, 
made according to Rowland's method,' with such a pitch as to 
measure wave-lengths directly in ten-millionths of a millimeter, a 
slight correction only being necessary, which depends upon the 
plate, and its position on the carriage of the engine. 

The basis of all the measurements is the Table of Standard 
Wave-Lengths.* As a means of determining the correction to 
be applied in each case, numerous standard lines selected from 
the Table are measured, in addition to those of the element 
which is being studied. Many of these standards occur on every 
plate, due to impurities in the carbon poles or in the material 
introduced into them. From the standards a correction-curve is 
plotted for each plate, due attention being given to the weights 
assigned in the Table, and to the character of the lines as affect- 
ing the accuracy of setting the cross-wires upon them. Correc- 
tions derived from the curve are then applied to the lines whose 
wave-lengths are required. The final value for the wave-length 
of any line is usually the weighted mean of several independent 
measurements, frequently upon different plates, and in the spectra 
of both first and second orders. 

The intensities given in the tables are estimated, in the case 
of the arc-spectra, upon an ascending scale from i to i ,000, in 
which I indicates a line so faint as to be just plainly visible on 
the plates. 

Notes marked (R) are by Professor Rowland. 

The intensities assigned to lines in the Sun, where these exist, 
are rather uncertain estimates, in which as a means of reference 
D, is taken as 300, and the Ni line at 5893.1, at 60. 

' See Emyclofiadia Britannica, Art. " Screw." 

•Rowland: "A New Table of Standard Wave - Lengths." A, and A. la, 
3ai» 1893- 



i6 



H. A. ROWLAND AND R. R. TATNALL 



BORON. 
{Preliminary: — w.4, a too to 4400). 



Ware-Length 



2496.867 
2497.821 



Intensity 

and 
Character 



65 r' 
80 r 



Remarks 



Intensity 
in 
Sun 



Within the assigned limits, these two lines, forming a pair 
(given in the Table of Standard Wave-Lengths), appear to be 
the only representatives of the line-spectrum of boron, although 
the whole region is more or less densely filled with a band-spec- 
trum, probably due to a compound such as boracic acid. 

BERYLLIUM. 
(Preliminary; — f».-/. 2too to 4600), 





Intensity 




Intensity 


Wave-Length 


and 


Remarks 


IB 




Character 




Sun 


2175.072 • 


2 






2348.697 


50 r 


Possibly double, with components 0.346 apart. 




2350.855 


7 






2494.532 


40/ 8 






2494.960 






2650.414 
2651.042 


45/ 8 






2898.352 • 


I 






2986. 187 • 


1} 






2986.546- 






3130.546 


Z} 


Coincides with a fine line in the Sun (R). 


4 ' 


3 131. 194 


Coincides with one-half of a broader solar line. 


5 n 


3321.219 


451 
45/ 


The latter is therefore probably double (R). 




3321.487 






3367.719* 


3 n 






4572.869 


45 « 







■ r indicates reversed, 

* Possibly not due to beryllium. 

3 On this scale the standard Zr line X 3129.882 has the intensity 3. 

r indicates reversed. 

d indicates double, 

s indicates sharp, 

n indicates hazy or nebulous. 



ARC-SPECTRA OF BORON AND BERYLLIUM 1/ 

A few faint lines, difficult to identify, and indicated thus 
(*), do not certainly belong to beryllium. As a rule these 
do not occur in the second order spectrum. A careful compari- 
son with the spectra of elements likely to be associated with 
beryllium has shown that they do not belong to C, Si, Ca, Al, 
Fe, Mn, Mg, Sn, As, Ce. 

Johns Hopkins University, 
November 28, 1894. 



ON SOME ATTEMPTS TO PHOTOGRAPH THE SOLAR 
CORONA WITHOUT AN ECLIPSE, MADE AT THE 
MOUNT ETNA OBSERVATORY. 

By A. RiccO. 
I. 

With the Huggins Apparatus.* 

My first attempts to photograph the solar corona without an 
eclipse were made at Catania with the Huggins apparatus on the 
day of the partial solar eclipse of 1893, April 16.' I obtained an 
image of a brilliant halo surrounding the solar disk ; but since 
there was no interruption of the image such as the Moon should 
have caused, I was convinced that the image did not represent 
the solar corona. 

The results of these first experiments could not, however, be 
regarded as conclusive, for during the eclipse the Sun was only 
from 6° to 25° above the horizon. I therefore continued 
the experiments in the months of May and June of the same 
year. The method was the same as that previously employed, 
except that the small blackened metallic disk, which served to 
intercept the direct image of the Sun formed by the speculum 
metal mirror, was no longer used. Dr. Huggins had previously 
recognized the fact that the disk serves no useful purpose. A 
change in the size of the disk or its complete removal produces 
no perceptible difference in the coronal image obtained. 

The suppression of the occulting disk and the extreme brev- 
ity of the exposure rendered the use of a driving-clock wholly 
unnecessary. It sufficed to place the solar image in the center 
of the field immediately before the exposure. This was easily 
accomplished by making an image of the Sun, formed by a 
small hole pierced in a sheet of metal attached to the telescope, 

' For a description of this apparatus see Mem, Spettr, /fal.t 13, 108. 

*Mem. Spfttr. Ital., aa. 

18 



PHOTOGRAPHY OF THE SOLAR CORONA 1 9 

concentric with a circle drawn upon the screen which receives the 
image. 

Previous to June 1 2 I obtained about twenty photographs in 
which the Sun is surrounded by a halo of nearly equal extent in 
all directions, but so shaded as to exhibit faintly marked rifts 
and streamers, which recall the appearance of the solar corona 
during eclipses. The extent of the corona varies from a quarter 
to half a degree in the various plates, according to the expo- 
sure and the development. But the streamers are not clearly 
marked in any of these photographs, and the rifts do not 
approach near enough to the solar limb to indicate with any 
certainty a structure like that of the solar corona. 

At the following summer solstice I took the Huggins appa- 
ratus to the Observatory on Mount Etna (2950'"), where I made 
a second series of photographs on bromide plates (Lumi^re and 
Dringoli) and chloride plates (Dringoli), thirty-two in all. On 
most of the days of observation at this altitude, during this 
season, the sky was so pure that by occulting the Sun by the 
edge of the opening in the dome the blueness persisted up to 
the solar limb, and there was no trace of a brilliant atmospheric 
halo. 

On this occasion, on account of a lack of snow or water at 
the Observatory, I had to postpone the development of the 
plates until my return to Catania, where it was carried out with 
the greatest care by Professor A. Mascari. 

We were at first greatly surprised and delighted to see on 
six of the photographs great curved streamers bearing a remark- 
able resemblance to the solar corona at the time of its greatest 
development. But the results were so unexpectedly good that 
our suspicions were aroused. It was soon noticed that all of 
the streamers were concave toward the axis of rotation of the 
exposing shutter, which turns in front of the sensitive plate at a 
short distance from it. The arcs formed by the rays, although 
irregular and interlacing, were nearly concentric with the axis. 
Moreover, when we had determined the direction of the solar 
axis, we saw that the streamers (which always corresponded 



20 A. RICCd 

closely in direction with the horizontal side of the plate) had 
their greatest extent before and after noon in various positions 
between the solar axis and equator. I finally concluded that the 
streamers might be produced by reflection of the very brilliant 
light of the solar image from the edge of the slit of the shutter. 
I recollected that I had noticed with surprise that the image fall- 
ing on the slit gave rise to metallic reflections, although the slit- 
faces were blackened. It therefore seemed probable that the 
black had been rubbed away from some part of the edges by 
the frequent opening and closing of the slit, and that the bril- 
liant reflections had fallen upon the sensitive plate. 

On my return to the Etna Observatory in July and August I 
made a third series of photographs, taking care to blacken the 
edges of the slit before each exposure with lampblack mixed 
in alcohol. In order to determine whether the streamers had 
any connection with the shutter, I varied its inclination consid- 
erably for the different exposures. I also made some photo- 
graphs with the slit-faces moved back so as to leave a circular 
opening o".07 in diameter. Still others were made with a sliding 
shutter operated by hand, which I made myself in the little work- 
shop of the Observatory on Etna. 

I obtained in this way twenty-two negatives, but on none of 
them were the streamers visible. The photographs made with a 
slit-width of less than o"*.04 in most cases showed the ordinary 
halo, the clearly marked streamers of which exhibited a polyg- 
onal or stellate form. 

The negatives made with greater slit-widths or with the slid- 
ing shutter showed a reversed solar image, and the halo, which 
was often visible in spite of the density of the sky, was of the 
ordinary form. 

In September I made a fourth series of fourteen photographs. 
By scraping the edges of the slit so as to expose the metal in 
several places, I obtained by reflection on the plates arcs some- 
what more regular and more exactly concentric with the axis 
of the shutter and less dense near the image of the solar disk, 
but nevertheless closely resembling the streamers which I 



PHOTOGRAPHY OF THE SOLAR CORONA 21 

obtained in the second series. There was thus no further doubt 
as to their purely instrumental origin. 

On July 13, near the solstice, and again on August 14 of the 
present year (1894), I made a fifth and sixth series of photo- 
graphs of the corona, and also some of the Moon for comparison. 
Part of the negatives were made on very sensitive Cramer plates, 
for which I am indebted to the kindness of Professor Hale. 
The greater part of the twenty-four photographs of the corona 
turned out very well, but in all cases the characteristic structure 
of the solar corona as seen during eclipses was absent. 

The difference between the photographs of the corona made 
at Catania and those made on Mount Etna is simply a variation 
of intensity and extent, such as might result from differences 
in the exposure or development. 

The photographs of the Moon, made on the less sensitive 
chloride and bromide Dringoli plates, showed nothing, even when 
the slit was wide open. With the more sensitive Cramer plates I 
obtained faint traces of the Moon in its last quarter an hour 
before sunrise, with slit-widths of o°*.o2 and o".o4. A negative 
of medium density was obtained with an exposure of 4*, twenty 
minutes after sunrise. 

If the corona was really obtained in the above experiments 
we may conclude that it has a stronger photographic effect than 
the Moon. 

It follows from photographs made by the astronomers of the 
Lick Observatory' that, if we take as unity the actinic intensity 
of the light of a Carcel lamp shining upon a surface at a distance 
of one meter through a hole one millimeter in diameter, the 
intrinsic actinic intensities of the brightest parts of the corona 
were: 

At the eclipse of August, 1886, - - 0.031 
At the eclipse of January, 1889, - - - 0.079 
At the eclipse of December, 1889, - - 0.029 

Expressed in the same unit, the intrinsic actinic intensity of 
the full Moon is 1.66, and that of the sky i'' from the Sun is 

' Keport of the Observations of the Total Solar Eclipse^ December^ i88g, p. 14. 



22 A. RICCO 

40.00. If we call the brightness of the corona 0.08, we have the 
ratio of the brightness of the corona : Moon : sky= 1:21: 500. 

Thus even when we take the corona at its brightest, and bear 
in mind the great difficulty of the measurements, we must admit 
that the light of the solar corona is always much fainter than 
that of the Moon, and far fainter than that reflected by the sky 
near the Sun. This conclusion is supported by the fact that at 
several eclipses the solar corona gave no sensible shadow, or at 
best a very faint one, while the Moon and the diffuse light of 
the sky gave a very strong one. 

As in my photographs the photographic action of the halo 
was far greater than that of the Moon, we must conclude that it 
does not represent the true solar corona, but rather the diffuse 
light of our atmosphere around the Sun, which is considerably 
brighter than that of the Moon. This is true even at an altitude 
of nearly 3000 meters, although in this case the atmospheric 
illumination must be much less brilliant. This is easily evident 
to the eye; for under favorable conditions no bright halo is 
seen, although it is always visible to an observer at the sea-level. 
By photography, however, it has been found that the halo, which 
is sometimes invisible from the summit of Etna, is still quite 
bright enough to make a strong impression upon sensitive plates. 

It is probable that part of the light of the halo results from 
the diffusion of direct sunlight from the imperfectly polished 
surface of the metallic mirror. 

II. 

WITH THE HALE APPARATUS.' 

In 1893 Professor Hale, Director of the Kenwood Obser- 
vatory of the University of Chicago, expressed to Professor 
Tacchini and me a desire to repeat upon Mount Etna the exper- 
iments in photographing the solar corona that he had been 
unable to carry to a successful conclusion on Pike's Peak, on 
account of the smoke from great fires in the surrounding forests. 
We assured Professor Hale that we would be very happy to 

'For a full description of this apparatus see Hale, A, and A, I3, 681. 



PHOTOGRAPHY OF THE SOLAR CORONA 23 

place at his disposal everything at the Etna Observatory that 
could be of service to him, and we subsequently agreed to ascend 
Mount Etna together at the time of the next summer solstice. 
During our stay on the mountain he was to carry on his experi- 
ments with the spectroheliograph, while I continued mine with 
the Huggins apparatus. 

We were so much retarded by the delay of Mr. Otto Toepfer, 
of Potsdam, in sending the spectroheliograph that we had to 
defer our expedition until July 7. Unfortunately there was just 
at this time a certain increase in the activity of the central crater, 
which was probably the prelude of the disastrous earthquakes on 
the eastern slope of the volcano which occurred on the 7th and 
8th of Augtist. The volumes of smoke rising from the great 
crater were carried by the prevailing northwest wind over the 
Observatory. There was also a serious difficulty in the fact that 
the sulphurous vapors mixed with the smoke tarnished the spec- 
ulum metal mirror of the apparatus. A mirror, made by depos- 
iting a film of platinum and gold upon glass, was to have been 
sent from Berlin, but it was not finished on account of insuper- 
able difficulties encountered in the process. As Professor Hale 
could not further prolong his stay, in the hope of enjoying 
atmospheric conditions comparable with those of my previous 
experience, he decided to leave the apparatus with me, in 
order that I might make use of the first favorable conditions to 
continue the experiments. 

While I greatly regretted that Professor Hale had had no 
opportunity to try his method of photographing the corona, I 
was much pleased by his confidence in me, and anticipated with 
pleasure the ex{>eriments I hoped to make with the s{>ectrohelio- 
graph on my next ascent of Mount Etna. 

On July 14 the mirror was removed, and the entire spectro- 
heliograph was carefully wrapped with cloth and paper. We 
then descended to Catania. 

When I returned to the Etna Observatory on August 10, I 
found the spectroheliograph badly oxidized, in spite of its many 
coverings. Several days were required to clean it and get it in 



24 A. RICCO 

order. As the mirror of gold and platinum had not arrived, 
I polished the speculum metal mirror as well as I could with 
absorbent cotton, wet with distilled water and alcohol. The 
entire mirroi cell was then carefully painted with lampblack 
mixed in alcohol. 

While awaiting perfect atmospheric conditions, I made a 
number of experiments in photographing the Sun's disk and the 
spectra of the Sun and sky. On August 24 the central crater 
emitted but little smoke, and the wind blowing from the east 
carried it far from the Observatory. The sky was of a beautiful 
blue in the zenith, and even near the limb of the Sun, though 
here the blue was lighter. I commenced at once to make photo- 
graphs of the corona, which I developed immediately with the 
solutions left by Professor Hale. In the six negatives obtained 
between noon and 3 p. m. on this day, the shadow or silhouette 
of the disk which covers the fixed image of the Sun is sur- 
rounded by a clearly marked halo, which stands out well from 
the background of the sky, but is not very intense. This 
halo has a nearly uniform diameter, and the outer boundary is 
nearly circular and equally shaded in all directions. There is 
no visible trace of coronal structure. 

The 25th of August was a cloudy day. On the 26th the sky 
was clear, but not so blue as on the 24th ; a faint, whitish halo 
was seen surrounding the Sun when the direct light of the disk 
was occulted by the side of the opening in the dome. Never- 
theless I made a second series of nine photographs of the corona, 
as I noticed that the purity of the sky gradually increased while 
I was at work. As arrangements had been made to leave the 
Observatory on this day, I did not have time to develop the 
plates immediately. They were developed two days later at 
Catania with the solutions used in our photographic work on the 
Carte du CieL The solutions were diluted at first, and strength- 
ened during the development, in order to render it uniform and 
complete in spite of the great heat of the season. But even with 
this precaution the negatives were fogged a little, perhaps on 
account of the high temperature, or possibly because the devel- 



PHOTOGRAPHY OF THE SOLAR CORONA 2$ 

oper employed was not so well suited for Schleussner plates as 
that used by Professor Hale. 

In this series also the halo has a regular form. There is no 
sensible difference between the results of the two series, although 
they were made under different atmospheric conditions. 

It is desirable to add some facts relative to the time of 
exposure of the negatives made with the spectroheliograph. In 
his preliminary experiments on Mount Etna Professor Hale 
found that the region of the K line in the spectrum of the 
Moon required an exposure of about 40 seconds with the 
Schleussner plates used. If in photographing the whole disk 
(11"" in diameter) the second slit is o"".2 wide, an exposure of 
^ X 40 seconds = 37 minutes would be necessary ; 1. e,, the slit 
should move 1 1"*" in 37 minutes. I found the brightness of the 
sky to be so great that this exposure was far too long to give 
the corona. The photographs on which the halo was most 
distinctly visible on a faint background of 3ky had exposures of 
only from 3 to 9 minutes for the whole run of the spectrohelio- 
graph (about 51""). These correspond to exposures of from 40 
to 120 seconds for the solar diameter. On the plates obtained 
with longer exposures the halo is lost in the dark background of 
the sky. The irregularities in the motion of the carriage are 
also more plainly evident, as at low speeds the motion is not so 
exactly controlled by the clepsydra. 

Thus, even if we employ the spectroscopic method, we obtain 
halos much brighter than the Moon (20 to 60 times), while 
visual observations show the Moon and the corona to be of 
nearly equal brightness, and photometric observations indicate 
that the Moon is the brighter. 

We thus reach the same conclusion as before. The halo 
obtained does not represent the true corona, but more prob- 
ably a bright atmospheric illumination surrounding the Sun, 
although the photographs are made with a dark line, in which 
the light of the atmosphere is lacking (or more exactly, is rela- 
tively feeble). It is difficult to believe that this halo is caused 
simply by diffusion of the light of the Sun's disk in the lenses 



26 A. RICCO 

and prisms of the apparatus, for when the first slit passes over 
the occulting screen it receives only the light from the regions 
outside of the photosphere. During the passage of the first slit 
there is certainly some diffusion throughout all, or nearly all, of 
its length. In fact, on four plates made with the slit moving 
rapidly across the image with the occulting screen removed, the 
sky is dark at the upper and lower limb of the Sun, while at 
those points where the slit in its motion becomes tangent to the 
limb, the sky is perfectly clear. 

Finally, it may be that the halo is partly caused by the 
imperfect polish of the mirror, especially in the case of this 
apparatus, the mirror of which had been tarnished by the 
sulphurous fumes from the central crater of Etna. In order to 
learn what share this cause may have had in producing the halo, 
it would have been desirable to photograph a very brilliant globe 
or disk, to see whether a halo would have been formed. Unfor- 
tunately I had neither the time nor the means to make these 
experiments on Mount Etna. 

OSSERVATORIO DI CaTANIA, 

September, 1894. 



PJ.ATE III 




Kk;. 1 




I-'IG. 2 




Fig. 3 



PHOTOGRAPHS OF THE VARIABLE STAR O. A, i6,i2i AND ITS 
SPECTRUM, TAKEN AT AREQUIPA. PERU 



DISCOVERY OF VARIABLE STARS FROM THEIR 
PHOTOGRAPHIC SPECTRA. 

By Edward C. Pickering. 

An illustration of the method employed at the Harvard Col- 
lege Observatory, during the last five years, for discovering 
variable stars of long period from their photographic spectra is 
shown in Plate III. The spectra of a large part of this class of 
variables are of the third type, and when near maximum the 
hydrogen lines are bright. With perhaps a single exception, no 
star has yet been found having this class of spectrum which 
has not proved to be variable. The Henry Draper Memorial 
furnishes every year photographs of the spectra of many thou- 
sands of stars. From an examination of these spectra Mrs. 
Fleming has discovered thirty-four new variable stars, and has 
shown that sixty-five known variables have a similar spectrum. 
No star has been assumed to be variable from its spectrum, but 
in each case, on the examination of photographic charts of the 
region, taken on different days, marked changes in brightness 
have been found. The variation has also, in every case, been 
confirmed by the writer before announcement of variability has 
been made. 

A photograph of the spectrum of 0. A. 16,121, which is 
Cord. DM.'-Tfi^ I3f626, in the constellation Scorpius, and whose 
approximate position for 1900 is R. A.= 16** 50".3., Decl.= 
— 30** 26', shows the hydrogen lines //y and HI bright on plate 
B 10,104. This photograph was taken with the Bache Telescope 
at Arequipa, Peru, on August 6, 1893, exposure 60". It is 
represented in Plate III, Fig. i, enlarged three times, so that 
the scale is i ' ssO**.! (nearly). The arrow indicates the spectrum 
of the variable, and also points to the hydrogen line //y, which is 
bright, and therefore appears dark in the print, which is a nega- 
tive. The line Hh is also seen to be bright and is about o***.55 
below Hy. The bright star of the first type 2°".6 to the left 

27 



28 EDWARD C. PICKERING 

of the variable is Cord. DM.—$o^ I3»564 mag. 6.8. The prin- 
cipal lines shown in the spectrum of this star are ff0t Hy, Hi, 
H€, K, Hi, Hii, etc. 

Enlargements of charts of the same region, and on the same 
scale, are given in Figs. 2 and 3. They are made from plates 
B6073 and B3802, taken on May 28, 1891, and July 13, 1889, 
with exposures of 10" each. In Fig. 2 the star is bright, 
while in Fig. 3 it is so faint as to be scarcely visible. Its 
maxima are represented by the formula J. D. 2,395,466 + 278 E. 
(See A, N. 135, 161). Fifteen other photographs of the region 
show the variable in different degrees of brightness from the 
magnitude 7.3 to 1 1.6. The evidence regarding the other thirty- 
three variables mentioned above as discovered by this method is 
equally conclusive. 

Harvard College Observatory, 
Cambridge, Mass., Dec. 14, 1894. 



PRELIMINARY TABLE OF SOLAR SPECTRUM 
WAVE-LENGTHS. I. 

By Henry A.Rowland. 

The following table of the lines of the solar spectrum has 
been in course of preparation at the Johns Hopkins University 
for many years in connection with an investigation of the spec- 
tra of the elements. The spectrum of every known element, 
except gallium (of which I have no specimen), has been photo- 
graphed in connection with the solar spectrum, and some of 
these plates have been measured. 

The whole solar spectrum has now been measured except 
the extreme ends, and the wave-lengths have been mostly 
reduced to my table of standards. Many of the lines, espe- 
cially the stronger ones, have been identified with respect to the 
substance producing them, but this must be a labor of years. 
Hence I have determined to publish the work so far as I have 
now proceeded, expecting to add to it and correct it for a term 
of years, until I can publish a standard list of the lines of the 
solar spectrum with all the elements to which they belong. 

The wave-length measurements have been made from my 
photographs of the solar spectrum, extending at present down 
to about wave-length 7200, and they will probably not be much 
changed in the future. The figures in the table refer to the 
wave-lengths in air at 20 ''C and 76'" of mercury, as they are 
based upon the table of standards. 

The intensities of the solar lines go from i, a line just 
clearly visible on my map, up to 1000 for the H and K lines. 
Below I the lines in the order of faintness proceed from o to 
0000, indicating lines more and more difficult to see. 

The ordinary scale from i to 10 or from i to 6 is far too 
limited for the spectral lines, especially for the metallic spectra ; 
I to 1000 is hardly great enough for the enormous difference in 
intensity. The small range, i to 10, ordinarily used gives an 

29 



30 HENRY A. ROWLAND 

entirely wrong idea to the worker in this subject, and many 
books with spectroscopic theories might have been saved by 
using a scale from i to looo. 

The expenses of this work have been partly borne by appro- 
priations from time to time by the Bache fund of the National 
Academy of Sciences, the Rumford fund of the American 
Academy of Sciences, and the Bruce fund of Harvard College 
Observatory. 

The measurement and calculation of wave-lengths have been 
made most carefully by Mr. L. E. Jewell. 

My thanks are due to very many friends for assistance in 
procuring pure and rare elements, etc. 

The symbols used are as follows : A line that is not clearly 
defined, or that is much weaker than it should be for .a line of 
its breadth, is indicated by an N (such" a line may generally be 
considered as composed of two or more lines too close together 
to be separated). Two lines close enough together to be con- 
sidered double are indicated by a d, and a double whose com- 
ponents are very difficult to separate is indicated by d ? ; like- 
wise three lines so close together as to be considered a triplet 
are indicated by a t. 

Where several lines are connected by a large brace there 
is generally shading extending from the center of some strong 
line to other lines at the extremities of the braces. This is 
the case with the stronger lines of iron and some other ele- 
ments. 

In the column devoted to the identification of the solar 
lines an interrogation mark indicates that the identification of 
the solar line with the element given is uncertain. Where two 
or more elements are given, the solar line is compound. The 
order in which they are given indicates the portion of the line 
due to each element, as follows : C-Fe-Cr. However, it is not 
always possible to correctly assign the exact positions, and 
consequently there are probably many errors in the positions 
assigned. Where the solar line is too strong to be due entirely 
to the element with which it is identified, it is represented thus : 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS Zl 

— Fe, and indicates that the iron line is coincident with the red 
side of the solar line, the origin of the rest of the line being 
unknown. If, as far as can be determined from the plates, two 
metallic lines coincide exactly with the same portion of a solar 
line, this is shown by using a comma instead of a dash. Thus : 
Fe, Cr. In some cases when a double line is particularly difficult 
to separate, measurements are given on the two components and 
also on the line unresolved. This last measurement is placed in 
parentheses between the other two measurements. Thus 

W.-L. INTENSITY. 

3738.454 1 3 ^ 

(3738.466) ■ 6 

3738.505 J 2^ 

means that there is a line at w,-L 3738.466 with the intensity 6; 
and that with good definition this line may be resolved into two 
components having the intensities 3 and 2 and the given wave- 
lengths. 

Johns Hopkins University, 
Baltimore, Dec, 1894. 



32 



HENRY A. ROWLAND 







Intensity 






Intensity 


Wave-length 


Substance 


and 
Character 


Wave-length 


Substance 


and 

Character 


3722.071 


Fc 


3 


3729.096 


Ni 


Zd? 


3722.174 




2 


3729.214 


C? 


oooN 


3722.280 




00 


3729.481 







3722.377 




I 


3729.666 




00 


3722.518 






oN 


3729.865 




000 


3722.639 1 


Ni 




3 


3729.952 


Ti 


3 


(3722.692) ^8 






10 


3730.154 




00 


3722.729 J 


Ti-Fc 




6 


3730.283 




000 


3722.899 









3730.450 




I 


3722.987 






I 000 


3730.5 M 


Fc 


3 


3723.319 




000 


3730.625 


Co 


2 


3723.425 




00 


3730.732 




00 


3723.533 




00 


3730.898 


Ni 


I 


3723.651 




00 


3730.950 


Cr 


I 


3723.750 




I 


3731.093 


Fc 


3 


3723.827 




00 


373l..^OI 




00 


3723.985 







3731.403 


Co-Zr 





3724.053 




00 


3731.523 


Fc 


3 


3724233 


Ti 


I 


3731.763 




00 


3724.399 




000 


3731.869 




00 


3724.526 


Fe 


6 


3731.956 




000 


3724.7 «6 


Ti 


z 


3732.072 


Mn 





3724.884 




00 


3732.177 


Cr 


2 


3724.970 


Ni 


z 


3732.284 




00 


3725.090 




oNd? 


3732.356 




000 


3725.300 


Ti 


X 


3732.545 8 


Ti-Fc-Co 


6 


3725.447 







3732.776 


C 


oNd? 


3725.638 




3 


3732.894 




2 


3725.806 




do 


3733.028 




000 


3725.978 




000 


3733.128 




000 


3726.164 







3733.222 




000 


3726.206 




00 


3733.338 




I 


3726.558 




oooNd? 


3733.469 s 


Fc- 


7d? 


3726.806 


Co 





3733.635 


Co 


iN 


3726.983 




00 


3733.798 




000 


3727.061 


Fc-Mn 


4d? 


3733.910 






[00 


3727.167 







3733.984 






000 


3727.244 


Fe 


3 


3734.160 






00 N 


3727.488 




I 


3734.278 


Co 




I 


3727.590 







3734.428 






000 


3727.672 




I 


3734.608 






00 N 


3727.778^. 
3727.826 J* 


Fc 


W 


3734.679 






I 




3734.807 






oN 


3727.965 




2 


3735.014 s 


Fc 


- 


40 


3728.183 




I 


3735.135 






00 


3728.278 




00 


3735.261 






oN 


3728.474 




N 


3735.387 






000 


3728.544 




00 


3735.485 


Fc 




4 


3728.813 


Ti-Fc 


2 


3735.587 






000 


3729.004 


Mn 


00 


3735.694 




1^000 



TABLE OF SOLAR SPECTRUM WA VE-LENGTHS 



33 







Intensity 






Intensity 


Wcve-lcBgth 


SabMuioe 


and 

Character 


Wavelength 


Snbstanoe 


and 
Character 


3735.843 




oNd? 


3742.286 




I 


3736.041 


Co 


I 


3742.4X8 


c? 


00 


3736.107 







3742.707 




2 


3736.187 




00 


3742.763 


Fe 


3 


3736.321 






rooo 


3743.086 


Cr 


I 


3736.434 






000 


3743.269 




00 


3736.619 






00 


3743.358 







3736.734 






000 


3743.508 S 


Fe 


< 6 


3736.854 






000 


3743.626 


Ti 


2 


3736.9588 


Ni 




3 


3743726 


Cr 


I 


3737.059 s 


Ca-Mn 




5 


3743.921 




X 


3737.174 






00 


3744.029 


Cr 


2 


3737.281 s 


Fe 




30 


3744.138 




000 


3737.441 






iNd? 


3744.251 


Fc 


4 


3737.718 






ooN 


3744.303 




000 


3737.900 






00 


3744.513 




000 


3738.026 






00 


3744.634 


Cr 





3738.128 






000 


3744.697 


Ni 


I 


3738 21 1 






00 


3744.898 




000 


3738.281 






000 


3744.959 




000 


3738.454 


Fc 


■31 


3745.X9O 




000 


(3738.466) 




;r 


3745.279 




000 


3738.505 




3745.371 




00 


3738.652 




I 


3745.491 


Co 


2 


3738.773 




000 


3745.617 


Ti 


I 


3738.900 




00 


3745.717 s 


Fc 


r 8 


3738.946 




000 


3745.751 


Ni 


I 


3739.140 




00 


3746.058 8 


Fe 


6 


3739.260 


Fc 


2 


3746.191 


Ni 


lo 


3739.370 


Ni 


3 


3746.287 




000 


3739.467 


Fe 


I 


3746.387 




. I 


3739.674 


Fe 


3 


3746.511 




00 


3739.926 


Ni 


I 


3746.618 


Fc 


2 


3740.085 


Bi? 


000 


3746.717 


Mn 


I 


3740.205 


Fc 


2 


3746.864 




00 


3740.386 


Fe 


3 


3747.065 _ 


Fc 


W 


3740.477 







3747.147 




3740.605 




0000 


3747.369 







3740.672 




0000 


3747.492 




00 


3740.952 






3747.694 




I 


3741.025 




0000 


3747.864 




000 


3741.205 


Ti 


4 


3747.965 




00 


374X.339 




00 


3748.144 


Ti 


r\ 


3741.453 







3748.232 


Ti? 


ON 


3741.619 




I 


3748.4088 


Fc 


, 10 


374«.70i 







3748.549 





.1 


3741.791 


Ti 


4 


3748.650 




3741.973 


C 


000 


3748.749 







3742.043 


C 


00 


3748.821 


Cr 


I 


3742.219 


Fe 


I 


3748.943 




OOD 



34 



HENRY A, ROWLAND 



Wave-length 



Substanoe 



3749.049 

3749.110 

3749.195 

3749.388 

3749.509 

3749.631 S 

3749.764 

3749.884 

3>49.994 

3750.082 

3750.283 

3750.349 

3750.448 

3750.648 

3750.823 

3750.916 

3751.015 

3751.136 

3751.234 

3751.367 

3751-592 

3751.735 

3751.802 

3751.967 

3752.055 

3752.334 

3752.408 

3752.556 

3752.648 

3752.830 

3753.003 

3753.134 

3753.282 

3753.482 

3753667 

3753.732 

3753.893 

3754.009 

3754.265 

3754.367 

3754.481 

3754.647 \ s 

3754.719 i 

3754.866 

3755.015 

3755.147 

3755.275 

3755421 

3755.593 

3755.714 



Fe 
Cr 

Ni 



Fe 



C? 
Co 
C? 
C? 



Mn 



Fe 



Co 
Zr 
Fe 

C 
C 
Fe 



Ti 
Fe 

Fe-Ti 



C 
CoC 



Co 



Intensity 

and 
Character 



O 

2 
I 

oN 

00 N 

20 



000 

00 



000 

0000 

I 

000 N 

X 

00 

2 

000 

I 

oN 

00 N 

I 

00 

I 

000 

000 

000 

3 

000 

00 

4 

I 
2 
I 
000 

6 

00 

000 

000 

00 

00 N 

W 

000 

ooN 

000 

o 

00 

iN 

00 



fN 



Wave-length 



3755.863 
3755.964 
3756.080 
3756.213 s 

3756,405 
3756.480 

3756.705 

3756.791 

3757.081 

3757.212 

3757.304 

3757.444 

3757.508 

3757.597 

3757.824 

3757.950 

3758.099 

3758.173 

3758.269 

3758.375 s 

3758.456 

3758.577 

3758.736 

3758.863 

3758.967 

3759.096 

3759.215 

3759.297 

3759.447 

3759.613 

3759.725 

3759.829 

3759.940 

3760.032 

3760.196 

3760.364 

3760.531 

3760.679 

3760.844 

3761.070 

3761.206 

3761.464 

3761.568 

3761.695 

3761.830 

3762.012 

3762.349 

3762.448 

3762.495 

3762.616 



Subatanoe 



Fe 

C? 

C? 

Mn 

C- 

Fc 

Cr 
C? 
C? 
Fe 
Cr-Ti 



Fe 



Zr 

La-Fc 

Ti 



C? 
Fe 

C? 
Fe 



Ti 

Fe 

C 

C- 

Ti 



C- 
C- 



Intenslty 

and 
Character 



00 
00 
000 

3 

00 

00 

000 

00 

4 

000 

I 

o 

00 

2 

4 

rooN 



•?? 

o 


loN 
00 N 
00 N 
oN 
I 
2 

I2d? 
o 

X 

00 
00 
000 

5 
I 
ooN 

4 

00 

00 

I 

7 

2 

000 

iN 

3 

2 
I 
I 
000 



TABLE OF SOLAR SPECTRUM WA VE-LENGTHS 



35 



WsTC* length 



3762.758 
3762.89s 
3763.008 

3763.147 
3763.3'I3 
3763.426 

3763.514 

376^6x4 

3763.709 

3763.945 s 

3764.117 

3764.251 

3764.359 
3764.422 
3764.522 
3764.727 
3764.787 
3764.986 

3765.059 
3765.194 
3765.441 
3765.689 
3765.847 
3766.234 

3766.377 
3766.460 

3766.594 
3766.801 

3766.955 

3767.105 

3767.218 

3767.341 s 

3767.493 

3767.574 

3767.682 

3767.787 
3767.842 
3768.034 

3768.173 
3768.232 
3768.38s 
3768.544 
3768.799 
3768.871 
3769.157 
3769.454 
3769.603 
3769.792 
3769.861 
3769.752 



Fe 



C 
C- 
C 
C 
-C 
C 



Fc 
Fc 
Fe 
-C 
C 
C 
Fc 
Fc 



Fc 



C 
Mn 



Fe 

C 

C-Cr-Fe.( 



Cr 



Intensity 
and 

Character 



I 

000 

OON 

2 

000 N 

0000 

000 

00 

I 
\ 10 

oN 
Li 

00 

ooN 

00 

00 

00 

000 

00 

o 

6 

I 

2 

00 

000 N 

00 

3 

I 

I 

o 

8 

iN 

o 

00 



00 

00 

3 
000 

2 





I 

oN 

ooN 

3 

000 
00 
00 



Wave-len^ 



3770.132 s 

3770.308 

3770.446 

3770.553 

3770.671 

3770.739 

3770.859 

3770.992 

3771. 116 

3771258 

3771.418 

3771.471 

3771.636 

3771.798 

3771.956 

3772.1 XI 

3772.250 

3772.330 

3772.524 

3772.673 

3772.730 

3772.918 

3773.070 

3773-345 

3773.503 

3773609 

3773.695 

3773.803 

3774.029 

3774.170 

3774.247 

3774.357 

3774-473 s 

3774.650 

3774.791 

3774.971 

3775-137 

3775.342 

3775.431 

3775.562 

3775.717 

3775.849 

3775.997 

3776.090 

3776.198 

3776.337 

3776.473 

3776.600 

3776.698 

3776.830 



Substance 



Fc 

C 

C-Fe 



C 
C 

Ti-C 
C 

-C 
C 

Ni 



C 
C 

c 
c 

Fc 

C 
C 

Y 

C 

Ti-C 

Fc 

C 
C 

Ni 
Tl 

C 
Ti 
C 

Fc 
Mn 



Intensity 

and 
Character 



M 



4 

00 

2 

2 

000 

I 

oN 

000 

2 

00 

000 

00 

I 

2 

00 

000 

oN 

00 

000 

2 

I 

000 

oNd? 

ooN 

I 

00 

00 

3 

00 

00 

00 

0000 

3 

000 

I 

4 
000 

00 
00 
000 

7 
000 

I 

00 
2 

oN 
000 N 

3 

I 
000 



'Haze» which is perhaps due to Hydrogen. 



36 



HENRY A. ROWLAND 







InleiMity 






Inlemity 


W«««-leagth 


Sabttanoe 


and 

dutnictcr 


Wcve-lencth 


Sbbbhikib 


and 
Chancier 


3776.977 




oooN 


3783.224 




000 


3777-057 




000 


3783.328 




00 


3777.210 


Fc-C 


2 


3783.483 




2 


3777.370 




000 


3783.601 


c 


00 


3777.470 




I 


3783.674 s 


Ni 


6 


3777.593 


Fc 


3 


3783.954 


C 


00 


3777.701 


C 


000 


3784.035 


c 


00 


3777.809 


C 


00 


3784.391 







3777.897 


C 


00 


3784.511 


c 


00 


3777.982 


c 





3784.641 


c- 





3778.075 


c 





3784.814 




000 N 


3778.203 


Ni 


2 


3784.965 




oooN 


3778.301 







3785.152 




000 


3778.463 


Fe 


3 


3785.223 




000 


3778.652 


Fe 


2 


3785.373 







3778.<J4i 


V-Fc-C 


3 


3785.457 


c 





3778.939 


C 


I 


3785.539 


c 





3779.049 




000 


3785.641 


c 


00 


377Q.165 




00 


3785.719 


c 


00 


3779.236 




0000 


3785.846 


Fc 


1 


3779.343 


C- 


I 


3785.929 




I 


3779.451 


C 


00 


3786.092 


Fc 


3 


3779.569 


Fc 


4 


3786.181 


Ti 




3779.657 


Fe 


2 


3786.314 


Fc 


id? 


3779.713 







3786.468 




I 


3779.871 


C 





3786.587 




1 


3779.989 


C 





37»6.66i 







3780.083 




000 


3786.820 


Fe 


5 


3780.223 


C 


000 


3786.983 




00 


3780.363 


C 


000 


3787.X04 




000 


3780.564 


C 


00 


3787.240 


Cr 


000 


3780.654 


C 


00 


3787.304 


Fc-C 


I 


3780.842 s 




3 


3787.379 


C 





3780.994 


C 


00 


3787.558 




000 


3781.128 




oooNd? 


3787.620 




000 


3781.330 • 


Fc 


3 


3787.713 




00 


3781.460 




00 


3787.852 




rooN 


3781.65s 




00 


3787.922 







3781.754 


c. 


I 


3788.046 s 


Fc 


" 9 


3781.818 


C 


00 


3788.189 




loo 


3781.935 
3782.078 




000 


3788.283 




00 


Fc 


2 


3788.353 




000 


3782.258 


Ti-Fc 


I 


3788.574 


C- 





3782.354 


C 





3788.666 


C 


00 


3782.45s 


C 


00 


3788.839 




2 


3782.592 


Fe 


2 


3788.953 


C 





i;l^:ill 


Fc 


I 
000 


3788.999 
3789.110 


Cr 
C 


I 




3782.987 




000 


3789.186 


C 





3783.135 


C 


000 


3789.319 


Fc 


3 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 



n 







Intensity 






Intensity 


Wsve- length 


Sabttanoe 


and 

Character 


Wave-length 


oiiiMuuioe 


and 
Character 


3789.439 







3795.44X 


c. 





3789.553 


Fc 


IN 


3795.5x3 


c 


00 


3789.637 







3795.593 







3789.719 


Fe 


I 


3795.680 


Fc 


I 


3789.867 







3795.880 


c 


00 


3789.991 




I 


3795.952 


c 


00 


3790.059 




000 


3796.040 







3790.238 


Fe 


5 


3796.154 







3790.362 


Mn 


2N 


3796.247 


c 





3790.471 


V? 


00 


3796.329 


c 





3790.585 


C 





3796.449 


c? 


00 


3790.629 


CrC 


I 


3796.53X 







3790.793 * 




I 


3796.635 







3790.910 


Fc-C 


2 


3796.943 




I 


3790.972 


La-Ca 


I 


3797032 


-c 


2 


3791.132 




000 


3797.1x5 


c 


00 


3791.246 


C 





3797-205 


C 


000 


3791.332 


C 


000 


3797.283 


Cr-C 





3791.5x7 


Cr-C 


I 


3797.387 


C 


00 


3791.645 


Fc-C 


2 


3797.597 




00 


3791.885 




I 


3797.659 


Fc 




r5 


3792.041 




000 N 


3797.860 


Cr 




I 


3792.216 


C 


00 


3797.99 X 


C 







3792.294 


Fe-Cr-C 


3 


3798.093 


Fc-C 




2 


3792.482 


Ni 


I 


3798.224 




t 


00 


3792.702 


C 





3798.306 






000 


3792.788 


c 





3798.396 


Mo 







3792.824 




2 


3798.486 




{ 


000 


3792.969 


Fc-C 


2 


3798.655 s 


Fc 


6 


3793.069 


C 


000 


3798.79X 


C 





3793.X28 




000 


3798.912 




000 


3793.262 




oooN 


3799.047 







3793.429 


Cr 


I 


3799.X59 




000 


3793.495 




I 


3799.272 


C 


00 


3793.622 


Fc 


2 


3799.386 


Mn-C 


iNd? 


3793.745 


Ni 


4 


3799.486 




00 


3793.846 


C 


000 


3799.586 




i.0 


3793921 


C 


00 


3799.693 8 


Fe 


3794.016 • 


Fe 


2 


3799.818 


C 


3794.107 




000 


3799.934 


V 


I 


3794.225 




000 


3800.046 







3794.3x3 




000 


3800.174 


C 





3794.485 


Fe-C 


4 


3800.256 


C 





3794.555 


C 





3800.457 


C? 


ON 


3794.679 







3800.683 


Mn 


I 


3794.753 


Cr 





3800.766 







3794.909 


La 


I 


3800.877 




00 


3795.032 


V 


fiN 


3800.992 







3795.147 s 


Fe 


^8 


3801.163 


Sn? 


000 


3795.292 




liN 


3801.251 








' Haze, which is perhaps due to Hydrogen. 



38 



HENRY A. ROWLAND 



Wave-length 



Sabttanoe 



3801.331 
3801.439 
3801.51 1 
3801.679 
3801.820 

3801.953 
3802.051 
3802.133 
3802.271 
3802.424 
3802.614 
3802.725 
3802.870 
3802.952 
5803,097 
?8o3.X4X 
3803.228 
3803.317 
3803.398 
3803.618 
3803.711 
3803.816 
3803.9x0 

3804.044 
3804.1 5 X s 

3804.237 
3804.317 
3804.424 
3804.484 
3804.621 
3804.752 
3804.836 

3804.934 
3805.070 
3805.256 

3805.337 
3805.486 s 

3805.589 
3805.669 
3805.884 
3805.989 
3806.103 
3806.255 
3806.357 
3806.511 
3806.586 
3806.7x1 
3806.865 

3807.012 
3807.X48 



C 

C 
-C 
Fc 
Fc 
Mn 



Fc 



C 
C 

C 
C 
C 

V? 



c 

Fc 
C 
C 



-C 

C 

FeCr-C 

C 
C 
Fc 
C 
C 



C 

Fc-C 

C 

C 

Mn-Fc 
V-Mn-C 



Intensity 

and 
Character 






oNd? 

3 

2 

00 

2 

00 

2 

00 

000 





o 



I 

00 

I 



00 

00 

000 

00 

3 

00 
00 

o 

000 

00 N 

I 
I 

2 

00 

00 

00 

6 

00 

00 

oN 

00 

000 



2 


o 

00 

8d? 


000 



Wave-length 



3807.293 
3807.425 
3807.539 
3807.681 
3807.831 

3807.914 
3808.076 
3808.223 
3808.274 
3808.423 
3808.659 
3808.770 
3808.873 
3809.189 
3809.305 
3809.552 
3809.633 
3809.724 
3809.834 
3809.894 

38x0.061 

3810.174 
3810.434 
3810.681 

38 1 0.76 X 

38x0.854 
38x0.901 
3811.047 
38XX.X8X 

3811.317 
.38x1.443 
3811.525 
38XX.667 
3811.787 
38IX.945 
38x2.033 
38x2.126 
38x2.205 
3812.340 
3812.389 
38x2.589 

3812.733 
38x2.813 
3813.000 
38x3.100 
3813.2x9 
38 x3.40s 
38x3.537 
3813.637 
38X3.78X 



Sabaianoe 



Ni 



Fc 
C 

Cr-C 

Co 

Ti-Ni-C 

Fc 

V 

Fc . 

C- 

C 

C 
MnC 

Mn 
Mn-C 

C 

C 



C 

C 

C 

Fc-C 



-C 
C 



Fc 

C 
C 
C 



Fc 
Fc 
C 
C 

v-c 

Fc 



Intensity 
and 

Character 



6 

00 

00 

6 

00 

000 

I 



I 


000 

3 

X 

o 

00 
00 

4 
o 
o 

00 

00 d? 
cod? 



o 

000 

3 



I 

000 

o 

000 N 
000 N 

2 
2 
o 
o 

000 

00 

000 

000 

000 

oN 

5 

2 
o 

2 
O 

2 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 



39 







Intensity 






Intensity 


Ware-length 


Substance 


and 
Character 


Wweleneth 


Substance 


and 
Character 


38X4.034 


Fc 


> 


3819.938 


C? 




00 


3814.070 




3820.042 






°ld 


3814.154 







3820.102 


C 




oi** 


3814-264 







3820.197 






00 N 


3814.386 




000 


3820.337 






oN 


3814.500 


C 


00 


3820.444 






oN 


3814.671 1 


Fc-C 


^1 


3820.586 s 


Fc-C 


' 


25 


(3814.698) \ 




8U 


3820.702 


C- 




I 


3814.738 J 


-C 


3J 


3820.797 


C? 







3814.796 


C- 


X 


3820.889 


C 




I 


3815.038 


C 





3820.950 


C 







3815.222 




00 N 


3821.017 









3815.352 




00 N 


3821.130 






.00 


3815.462 






I 


3821.328 s 


Fc 


4 


3815.572 


Cr 




I 


3821.636 




00 


3815.759 






oN 


3821.725 




00 


3815.987 s 


Fc 


, 


15 


3821.866 


C 


IN 


3816.252 


C? 




oN 


3821.981 


Fc 


4 


3816.332 


C? 







3822.077 


Ti 





3816^90 


Fe-Co 




.3 


3822.157 


V 





3816.610 


Co 


I 


3822.250 




000 


3816.779 


C 


ooN 


3822.406 


C 





3816.887 


Mn 


I 


3822.470 


C 





38^6.998 




0000 


3822.557 




000 


3817.059 




I 


3822.785 




00 


3817.II4 




000 


3822.924 




000 


3817.198 


C 


00 


38^2.996 


V 


Id? 


3817.290 


C 


000 


3823.092 


C? 


00 


3817.523 


Co-Ti-C 





3823.163 


C 





3817.602 




00 


3823.228 


C 





3817.725 




00 


3823.352 







3817.786 


Fc-C 


3 


3823.493 




ooN 


38X7.877 


C 





3823.653 s 


Mn-Cr 


4 


3817.985 


Ti-Cr-C 





3823.893 


C 





3818.089 




000 


3823.953 


C 





3818.225 




000 


3824.028 


Mn 


I 


3818.337 


C 


00 


3824.139 


C? 


00 


3818.378 


VC 


I 


3824.216 


Fc 


I 


38x8.487 




I 


3824.373 




000 


3818.613 


Cr 





3824.441 




f^ 


38X8.759 


C 


I 


3824.591 


Fc 


V 


3818.89X 


C? 


000 


3824.711 




Y 


3819.O3X 




000 


3824.780 




looo 


3819.197 


C 


iNd? 


3824.887 




00 


3819.346 


C 


I 


3824.936 







3819.412 


C 


1 


3825.057 




1 


3819.521 


C 


00 


3825.183 




000 


3819.637 




2 


3825.256 




00 


38x9.715 


Cr 





3825.373 


C 





38i9U{29 


C- 


iNd? 


3825.448 


C 






40 



HENRY A. ROWLAND 



Wave -length 



3825.543 

3825.736 

3825.820 

3826.027 S 

3826.163 

3826.229 

3826.343 

3826.449 

3826.555 

3826.636 

3826.761 

3826.843 
3826.905 
3826.988 
3827.096 
3827.220 
3827.348 
3827.435 
3827.519 
3827.622 

3827.714 
3827.828 
3827.980 S 

3828.155 
3828.296 
3828.360 

3828.539 
3828.646 
3828.702 

3828.795 
3828.971 
3829.108 
3829.195 
3829.284 
3829.386 

3829.501 s 
3829.617 
3829.728 
3829.822 

3829.909 
3830.037 
3830.211 

3830.447 
3830.513 
3830.627 

3830.745 
3830.801 
3830.896 
3831.002 
3831.174 



C? 
Fc 



C? 
C? 



C 
C 
-C 
C 



Fe 

Ti-C 
C 
C 
V 



Fc 

Mg 
Ti-C 
Fc-C 
Ti-C 



C 

c 

Fc 
Fc 
C- 



Intensity 

and 
Character 



{ 



2 

I N 

I N 

20 

oN 



o 

o 

I 

000 

1 N 



00 

2 



o 

00 

I 
o 

000 

2 

iN 

8 

iN 

I 



00 

o 



000 

00 

000 

000 

I 

00 

10 

2 

o 

I 

X 
000 

o 
000 



o 
o 


2 
2 

3d 



Ware- length 



3831.334 
3831.517 
3831.657 
3831.837 
3832.025 

3832.171 
3832.303 

3832.450 S 

3832.647 
3832.790 
3832.890 
3833.026 
3833.156 
3833.221 
3833.348 
3833.458 
3833.628 

3833.744 
3833.836 
3833.916 
3834.006 
3834.096 
3834.189 
3834.364 
3834.506 

3834.614 
3834.609 
3834.762 
3834.869 
3834.978 
3835.022 
3835.176 
3835.298 
3835.342 
3835.509 
3835.689 
3835.862 
3836.116 

3836.229 s 

3836.337 
3836.476 

3836.639 \ s 

3836.689 f ^ 

3836.808 

3836.905 

3837.059 

3837.277 

3837.404 

3837.559 

3837.768 



Snbatanoe 



Intentity 

and 
Character 



Ni 
C 



Mg 
C 



C 

c- 

Fc 

C 

C 
Mn-C 



Fe 
Mn 

C 
C 



C 
C 
C 
C 
-C 



C 
Fe 
C 
C 



C 
C 



00 Nd? 
ooN 
000 
6 
'0 

o 

15 



000 
UN 
o 
I 
o 

4 

oooN 

o 

o 

o 

'3 

000 

oN 
10 

4 



o 

Loco 



ooN 
000 
ooN 
o 
o 

Id? 
oNd? 
000 d 
^000 

3 
000 

3 

\V 

000 
I 

I 

2 

000 

oNd? 

Id? 



' Haze, which is perhaps due to Hydrogen. 



TABLE OF SOLAR SPECTRUM^ WAVE-LENGTHS 



41 







IiMciisity 






Intensity 


Wm>kacth 


SnMtsnos 


and 

Cosradttr 


Wave-length 


Subetanoe 


•ud 
Character 


3837.961 


C 


fo 


3844.714 







3838.035 


c 







3844.861 




ooN 


3838.188 






ON 


3845.023 







3838.345 






IN 


3845.149 


C 


I 


3838.435 s 


Mg-C 




25 


3845.310 


Fe 


3 


3838.675 






oN 


3845.358 




I 


3838.888 


c 




iNd? 


3845.461 


C 


000 


3839.135 









3845.606 


Co-C 


8d? 


3839.375 


c 


I 


3845.729 




000 


3839.405 


Fe 


3 


3845.837 




I 


3839.582 







3845.949 


C 


00 Nd? 


3839.762 


Fc 


2 


3846.131 


C 


2 


3839.932 


Mn-C 


2 


3846.421 




I 


3839.987 


C 





3846.554 


Fc 


3 


3840.067 




00 


3846.666 




00 


3840.339 


C? 







3846.777 


C 


I 


3840.340 


Mn 







3846,814 


C 


I 


3840.440 


V 




ON 


3846.943 


Fe 


5 


3840.580 8 


FcC 




8 


3847.087 


C 


I 


3840.730 
3840.893 






ON 


3847.121 


C 


00 


V 




1 N 


3847.261 




000 N 


3841.034 


C 




oNd? 


3847.394 


C 


ooN 


3841.195 


Fc-Mn 




10 


3847.477 


V 


00 


3841.3*7 









3847.567 




00 


3841.430 


Cr 




I 


3847.654 




000 


3841.486 




00 


3847.827 




000 


3841.595 


Co 





3847.961 


C 


X 


384X.73O 




oooN 


3848.006 


C 


I 


3841.86a 


C 


3d? 


3848.100 


C 





3841.959 
3843.083 


C 





3848.186 




00 


C 





3848.249 




00 


3843.191 


Co 


3 


3848.339 


C 


00 


3843.334 


•C? 


od? 


3848.433 




3 


3843.500 




000 


3848.580 




000 


3843.587 


C 


00 


3848.667 




00 


3843.779 


C 





3848.745 




00 


ssust 




000 


3848.840 




00 




1 


3848.979 


C 


I N 


3843.137 


C 


3 


3849.140 


LaC 


3d? 


3843.195 


Fc-C 


3 


3849 248 




000 




Fe 
C 


^N 


3849.400 
3849.501 


C? 


000 

I N 


3843.903 




000 


3849.675 
3849.888 




I 




3N 


C 


oN 




000 


3850.013 







3844.131 
3844.107 


Mn.C 


3 


3850.118 


Fe 


10 




0000 


3850.300 


C 


iNd? 


3844.367 




000 


3850.440 




00 


C 


4<l? 


3850.536 




00 


V 





3850.626 








42 



HENRY A. ROWLAND 







Intensity 






Intensity 


Wave-length 


Substance 


and 
Character 


Wave-length 


Substance 


and 
Cohaiactet 


3850.700 




00 


3856.674 







3850.781 


C? 





3856.802 


C? 


2N 


3850.962 


Fe 


4 


3856.955 




000 


3851-098 


Co? 





3857.063 


C? 





3851.220 




000 


3857.135 


C? 


000 


3851.306 




00 


3857.215 


C? 


000 


3851.427 


C 


2Nd? 


3857.288 


C?- 


I 


3851.580 




000 


3857.473 


C 





3851.672 


C 





3857.580 


C 


00 


385 '.733 


C 


00 


3857.805 


C? 


6d? 


3851.815 


C 


00 


3857.955 


C? 




3851.895 




00 


3858.033 


C? 




3851.993 


CoC 


00 


3858.146 


C? 


00 


3852.132 




000 


3858.262 




I N 


3852.245 


C 


00 


3858.442 


Ni 




3852.347 


C- 


I 


3858.606 


C 




3852.541 


C? 


2Nd? 


3858.642 


C 


000 


3852.714 


Fc 


4 


3858.722 


c 




3852.845 


C 


00 


3858.822 


c 


2N 


3852.899 


C 


000 


3859.000 




I N 


3853.045 


C 


00 


3859.052 


c? 




3853.184 




00 


3859.128 


c 


00 


3853.333 


C 


I d? 


3859.246 


c? 


00 


3853.477 


C 


ON 


3859.355 


Fe 


3 


3853.620 


C 


2d? 


3859.415 


C 


000 


3853.805 


C 


00 


3859.535 




000 


3853.872 


c. 





3859.568 


C? 


To 


3853.967 


C 


000 


3859.788 


C-Ni 





3854.040 


C 


00 


^859.876 


C 


> 00 


3854.191 


C 





3860.055 8 


Fc-C 


' 20 


3854.343 







3860.227 


C- 


oN 


3854.401 




00 


3860.354 


C? 


toN 


3854.507 




2 


3860.427 




00 


3854.707 


C 


2Nd? 


3860.566 


C 





3854.808 


C 





3860.630 


C 





3854.869 


c 





3860.767 


C-Ni 


3N 


3854.989 


c 


I 


3860.863 


C 


000 


3855.088 


c 


00 


3860.963 


C 





3855.259 


c? 


00 


3861.067 


C 


00 


3855450 


V 


2 


3861.158 


C- 





3855.547 




I 


3861.299 


C-Co-C 


4Nd? 


3855.721 1 


Fe-C 





3861.479 


C-Fc 


3 


(3855.749) \ 




5U 


3861.592 


C 


00 


3855.766 J 


C 


3j 


3861.681 


C 


I 


3855.989 


V 


4 


3861.734 


C 


2 


3856.107 




00 


3861.847 


C 


2N 


3856.161 




I 


3861.978 


C 


iN 


3856.282 




fooo 


3862.114 


C- 


iN 


3856.367 




i° 


3862.248 




000 


3856.524 8 


Fe 


18 


3862.361 




.000 



TABLE OF SOLAR SPECTRUM WA VE-LENGTHS 



43 







Inlensily 






Intensity 


Wave-leneth 


Substftiioe 


and 
Chancier 


Wave-length 


Subatanoe 


and 
Character 


3862^58 


C 


00 


3867.996 


C? 


000 


3862.541 


c 


000 


3868.060 


-C 


2 


3862.627 


c?. 


2 


3868.171 


C 


00 


3862.727 




I 


3868.261 


C 





3862.827 


c 


000 


3868.372 


-C 





3862.897 


c 


00 


3868wt5I 




000 


3862.962 







3868.539 


C 


I 


3863.041 


c 


00 


3868.625 


C 


00 


3863.113 


c 


00 


3868.700 


C 





3863.201 




I 


3868.785 


C 


00 


3863 341 




oooN 


3868.873 


C 


I 


3863.533 


c 


3N 


3868.941 


C 





3863.655 


c 


00 


3869.179 


C 





3863,734 


c 


00 


3869.305 


C- 


I 


3863JJ35 


c 


I 


3869.444 


C 


oNd? 


3863.888 


Fe 


3 


3869.533 


C- 


1 


3864.006 







3869.692 


Fe-C 


3 


3864.113 


C 





3869.745 


C 


I 


3864.246 


Mo-C 


I 


3869.805 


C 


1 


3864.4388 


C 


3 


3869.960 


C 


00 


3864.626 


C 


I 


3870.053 


C-Co 


iN 


3864.720 


C 


00 


3870.204 


C 





3864.802 


C 


00 


^870.289 


C- 


1 N 


3865.005 


V 


3Nd? 


3870.405 


C? 


00 


3865.134 


C 





3870.493 


C? 





3865.213 


c 


000 


3870.615 


C- 


I 


3865.282 


c? 


,3 


3870.685 


C 


oN 


3865.454 


c 


[0 


3870.797 


C 


° id 
ooi 


3865.554 


c 


J 


3870.848 


C 


3865.674 


Fe-C 


|7 


3870.932 


C 





3865.793 


C 


lo 


3871.018 


C 


I 


3866.046 


C? 


000 


3871.145 


C 





3866.122 


C? 


3Nd? 


3871.259 


C 





3866.238 


C 


00 


3871.356 


C- 


I 


3866.306 


C 


00 


3871.527 s 


C 


«2d? 


3866.380 


C 


00 


3871.693 


C 





3866.526 


C 


00 


3871.785 




00 


3866.577 


C- 


I 


3871.963 


Fe 


2 


3866.692 


C 





3872.035 







3866JJ54 







3872.202 


C- 


iN 


3866.9^ 


C- 


2 


3872.312 


C 


•0 


3867.118 


C 





3872.405 


C- 


iN 


3867.205 
3867.356 


C 





3872.639 


Fe 


6 


Fe-C 


3 


3872.859 


C 


I Nd? 


3867.449 


C 





3872.969 




00 


3867.520 


C 





3873.065 


-C 


2 


3867.573 







3873.224 


Co 


2 


3867.758 


CV 


I 


3873.267 




2 


3867.791 


C 





S873.333 




00 


3867.906 


C- 


I 


3873.427 








■ Beginning of the second head of " Cyanogen band." 



44 



HENRY A. ROWLAND 







Intensity 






Intensit, 


Wtve-length 


Sotetance 


and 

Chancier 


Wave-length 


SnbaiBDoe 


and 

Character 


3873.504 


C- 





3879.458 


C 





3873.636 


C 





3879.578 


C 


000 


3873.706 


C- 


I 


3879.716 


C 


I 


3873.903 


Fe 


4 


3879.796 


C 





3874.091 


Co-C 


4 


3879.851 


C 





3874.191 




I 


3879.986 


C 


00 


3874.258 


C 


00 


3880.105 


C 


I 


3874.328 


C- 





3880.175 


C 





3874.491 




000 Nd? 


3880.235 


C 





3874.651 




2 


3880.328 


-C 





3874.708 


-C 





3880.393 




0000 


3874.861 


C 


I 


3880.465 


C 




3874.911 


C 


2 


3880.532 


C 




3875.220 S 


V 


2 


3880.596 


C 




3875.425 


Ti-C 


2Nd? 


3880.684 


C 


000 


3875.513 


c 


I 


3880.815 


C 




3^75.681 




ooN 


3880.931 


C 




3875793 




00 


3881.038 


C 


000 


3875.920 


c- 


2 


3881.140 


C 




3876.019 


c 





3881.254 


C 




3876.083 


c 





3881.^46 


C 




3876.194 


Fc 


5 


3881.445 


C 




3876.448 


C 





3881.543 


C 




3876.556 


C 





3881.628 


C 


000 


3876.622 


C 





3881.729 


C 




3876.702 




000 


3881.825 


c 




3876.815 


Fc 


I 


3882.011 


C-Co 




3876.981 


Co-C 


4 


3882.118 


C- 




3877.121 


C 


4Nd? 


3882.224 


c 




3877.232 




00 


3882.309 


c 




3877.337 




I N 


3882.439 


c. 




3877.481 


C 


I 


3882.530 


c 




3877.587 


C 





3882.650 


c 




3877.646 


C 





3882.733 


c 




3877.745 




00 


3882.828 


c 




3877.845 




000 


3882.893 


c 




3877.972 


C 


fo 


3882.986 


c 




3878.072 


C 


00 


3883.033 


c 




3878.152 


Fc-C 


i8 


3883.133 


c 




3878.334 
3878.438 




[i 


3883.253 


c 




C- 






3883.339 


c 


.0 


3878.549 


C 




od? 


3883.426 { 

3883.533 r 


c- 


2 


3878.720 


Fe 




7Nd? 


c 


«iN 


3878.816 


CoFe 


, 




3883.568 s 




• 


3878.884 


C-Fe 






3883.690 




000 


3878.975 


C 






3883.778 s 


Cr 





3879.037 


C 






3884.236 




000 


3879.178 


C? 


ooN 


3884.361 




00 


3879.331 


C 


I 


3884.431 







3879.394 


C 





3884.518 


Fc 


2 



■First line in first head of "Cyanogen band." This line and the preceding one 
were measured together as a single line iKi the series of measurements upon which 
were based Rowland's Table of Standards. 

'Edge of "Cyanogen band." 



TABLE OF SOLAR SPECTRUM WA VE-LENGTHS 



45 



Wavc-kagth 



3884.579 
3884.748 
3884^12 
3884*986 
3885.101 
3885.207 
3885^90 

3885.364 
3885.426 

3885.657 
3885.797 
3885.895 
3886.005 

3886.073 
3886.205 

3886.295 
3886.434 s 

3886.568 
3886.942 
3887^)80 
3887.196 
3887.512 
3887.666 
3887.870 
3888.030 

3888.179 
3888.560 
3888.671 
3888.863 
3888.971 
3889.077 
3889.24s 
3889.374 
3889.498 
3889.675 

3889.810 

3889.986 
3890.069 

3890.222 

3890.336 
3890450 
3890.538 
3890.707 
3890.861 
3890.986 
3891.084 
3891.339 
3891.521 
3891.649 
1891.820 



Ca 
Fe 



Fc 

FeCr 

Co 

Fe 



Fe 

V 

Cr 

Fc 

Ti 

Fe 

FeMn 

Mn 
Ni 

V 
Fe-Zr 

Fe 



Intensity 
and 

Character 



1 



00 

I 
I 
o 

00 
00 

2 
2 
00 

4 

00 
I 
o 



ooN 
00 Nd? 

15 


L3 
00 

7 
oooN 

oooN 
000 N 
00 


'2 

5 

ooN 

2 



I 

iNd? 

Id? 

ooN 

2 



I 

000 

iN 

000 

2 

iN 

00 

3 

00 

iN 



iN 

000 



Wnve-lengtfa 



4.057 ^ 
4.165 ^ 

4.211) V 

4.241 J 



3S9I.9.S 
3892.069 
3892.153 
3892.373 
3892.450 
3892.590 
3892.698 
3892.873 
3893.033 
3893.124 

3893.213 
3893.352 
3893.451 
3893.542 
3893.600 
3893.743 
3893.932 
3894.057 
3894. 
(3894.' 
3894.' 
3894.355 
3894.511 
3894.630 
3894.768 
3894.850 
3895.119 
3895 224 
3895.304 
3895.377 
3895.470 
3895.583 
3895.719 
3895.803 
3895.931 
3896.279 

3896.385 
3896,500 
3896.608 
3896.671 
3896.759 
3896.917 
3897.119 
3897.210 
3897.336 
3897.482 

3897.596 s 

3897.785 
3897.915 
3898.032 



Fe 

Mn 
V-Fe 

Fe 



Fe 
Cr 

Co 



Mn 
Co 
Ce 

Ti 

Mn 

Fe 

V? 
Mn 



Zr 
Ce 



Fe 
Zr 

Fe 



Intensity 
and 

Character 





4 

000 

00 



00 

2 

000 

2 

I 

I 

000 

2 

4 


000 
000 

2 

i}. 

0000 
0000 

I N 

00 

o 

3 

I 
00 

2 
I 
3 



7 



o 

00 

00 

I 





oN 

000 

000 

000 

000 

2 





3 



' Hase, which is perhaps dac to Hydrogen. 



46 



HENRY A. ROWLAND 







Intensitr 






Inloiaily 


Wave -length 


SubtUunce 


and 
Character 




Sobstanoe 


and 

Character 


3898.151 


V 


5 


3905.017 




X 


3898.231 


Fe 


2 


3905.146 




X 


3898.414 




00 


3905.326 






2 


3898.531 


Mn 


2 


3905.497 






iN 


3898.645 


Ti 





3905.660 s 


Si 


- 


12 


3898.9x1 




000 N 


3905.816 






2N 


3899.015 




0000 


3905.906 






I N 


389917? 


Fe 


3 


3906.044 




3 


3899.277 




2 


3906.169 




00 


3899.463 




000 


3906.318 




000 


3899.530 


Mn 





3906.438 


Co 


2 


3899.701 


Mn 


foN 


3906.539 




000 


3899.850 


Fe 


J*^ 


3906.628 


Fe 


10 


3899.963 




] oN 


3906.763 




00 


3900.361 




lo 


3906.890 


Fe 


4 


3900.470 




000 


3907.099 




X 


3900.549 







3907.251 




oNd? 


3900.681 


Ti-Fc-Zr 


5 


3907.368 




00 


3900.797 




000 


3907.433 




000 


3900.907 




000 


3907.615 


Fe-Sc 


3d? 


3900.973 







3907.807 




X 


3901.1x4 


Ti 


I 


3907.910 




X 


3901.197 




000 


3908.077 


Fe 


5 


3901-297 




00 


3908.203 




000 


3901.474 




00 Nd 


3908.310 




0000 


3901.621 




00 


3908.410 




I 


3901.735 




2 


3908.546 







3901.878 




000 


3908.684 







3902.002 




3 


3908.821 




00 


3902.114 


Fe 


I 


3908.900 


Cr 


4 


3902.241 







3909.064 




I 


3902.399 


V 


3 ^ 


3909.211 




000 


3902.567 




I N 


3909.423 







3902.768 




2N 


3909.538 




0000 


3902.916 




f^ 


3909.638 




00 


3903.090 


Fe-Cr 


^10 


3909.802 


Fe 


4 


3903.216 




lo 


3909.863 




000 


3903.302 




I 


3909.976 


Fe 


5 


.3903.398 




2 


3910.079 


Co-Ca 


3Nd? 


3903.553 




000 


3910.21 1 




000 


3903.683 




00 Nd? 


39x0.348 




000 


3903.868 




000 


3910.469 




2 


3903.991 1 




*1 


39x0.615 







(3904.023) \ 




8 \a 


3910.670 




2 


3904.052 J 


Fe 


S) 


3910.802 







3904.213 


Co 





3910.984 


Fc-V 


4 


3904.467 




00 


39XX.135 




3 


3904.613 




000 


39 XX. 230 




00 


3904.767 







39x1.316 


'Nd 





3904.926 


Ti 


3 


39x1.444 




000 



' Nd is the symbol for Neodymium. 



PLATE IV 



□ □ 

X894, May 95, sh. om. a. m. Melbourne >894> M>y >6, sh. tsm. a. m. Melbourne 

Mean Time. Long. 114 . Power 980. Civil Time. Long, no . Power 980. 




1894, May 99, sh. som. Melbourne 
Civil Time. Ijong. 87 Power aSo. 





1894, May 30, 6h. ism. a. m. I^one. 83 . 1894, June 7, 6h. 40m. a. m. Mt-Umiime 

Power 330. Civil Time. I/ing. 10 . Pbwer 330. 

NORTH 

DRAWINGS OF MARS 

.Made with the Mcllwurne 4-foot Kertecior 

Bv FiKTRo Haracchi 



OBSERVATIONS OF MARS MADE IN MAY AND JUNE. 
1894, WITH THE MELBOURNE GREAT TELE- 
SCOPE.' 

By R. L. J. Ellery. 

May 25. — This is the first clear morning after May 21. 
Clear; heavy dew falling. Bad definition. Best power 280. 
The first general impression is as follows: 

Gibbosity on / side conspicuous. The South Polar Cap, is 
an elliptical area, dazzling white , with sharp contour, not a speck 
visible in it ; all along its northern side it is bordered with a 
dark narrow strip which is conspicuous. Faint dark patches 
of irregular, unrecognizable form extend from the border over 
a considerable portion of the southern half of the planet, 
a central marking, darker than the other, extending farther 
north. 

The / limb is very brilliant ; the terminator is dusky on the 
southern part and generally dull, but of sharp contour. All the 
bright parts of the planet are of a bright rose-orange. The mark- 
ings show no particular color, and look like pencil markings 
made on orange-tinted paper. 

Unsteadiness very persistent and troublesome. 

The sketch shows the appearance of the planet as seen by 
occasional glimpses of steadiness between 4.50 and 5.10. 

May 26, — Clear; frosty ; heavy dew falling. 

Image very unsteady. Best power 280. 

The northern border of the South Polar Cap does not seem so 
much curved as on May 25, the curvature being very slight this 
morning. The area, however, is about the same. Markings very 
faint and indistinct; they seem to change strangely, and it is diflS- 
cult to preserve any definite impression. 

Color and general character of image much the same as on 
May 25. 

'Communicated by Professor W. H. Pickering. 

47 



48 R. L. /. ELLERY 

Sketch, from impressions obtained by occasional moments of 
steadiness, from 5**20" to 5^30". 

May 2y and 28. — Overcast. 

May 2g. — Clear in parts. Thin moving banks of haze. 
Image extremely unsteady. Now and then there come moments 
of good definition, and a glimpse of the real image is obtained. 
The contour of the South Polar area is very sharp. Nothing to 
be seen within the area, which is always dazzling white. The 
coloring is the same as on previous nights. 

May JO, — Cloudy. The planet is only visible in occasional 
breaks. Generally unsteady and not quite clear ; but on two 
occasions I got a moment of very good definition and steadiness. 
There is a dash appearing on the s limb, near the /corner of the 
Polar Cap. This is the first marking seen on the area. 

May ji to June 6, — Overcast and raining. 

June f, — Partly cloudy. Clouds rolling on the surface as I 
watch it. The South Polar area is smudged on the / corner, as 
if the dash seen yesterday morning traveled to the opposite 
corner. 

Ju7ie 7 to 14, — Overcast and raining. No more observations 
made. 

General Remarks Applying to All the Observations, — The follow- 
ing limb and also the northern parts of the planet, were always 
very bright, and of a rose-orange color. The South Polar Cap 
was white both before and after sunrise. No change of color 
was observed in daylight except the orange became more rosy. 
The area on June 7 seemed smaller than on May 25. Mark- 
ings on sketches are drawn darker and slightly more pronounced 
than as seen originally. 



RECENT CHANGES IN THE SPECTRUM OF NOVA 

AURIGiE. 

By W. W. Campbell. 

Some interesting and significant changes have occurred 
recently in the spectrum of Nova Auriga, The intensities of at 
least two of the prominent bright lines have decreased very 
materially. They are the lines at X4360 and X 5750. The 
variations will be realized best if we tabulate, as below, the 
intensities of the principal lines as estimated in 1892, and of the 
same lines as estimated quite recently. 

Hy X43^ ^f^ X4960 X5010 X5750 

1892, Attgusl and September' 0. i 0.8 '13 10 i 

i894,May8 o.i 0.3 i 3 10 0.4 

1894, September 7 o.i 0.2 I 3 10 0.4 

1894, November 28 o-i 0.1 i 3 10 0.3 

It will be seen that the change in X 5750 is very decided, and 
that the change in X4360 is radical. The wave-lengths of both 
these lines were observed visually in August, 1892, with com- 
parative ease. The line X 4360 is now so faint that it can be 
seen only with great difficulty, and the line X 5750 is probably 
too faint to measure. 

On the spectrum photographs taken early in September, 1892, 
the line X4360 was by far the brightest line of all, being certainly 
eight times as bright as Hy. I have just secured another photo- 
graph of the spectrum (November 28), and it shows that 
A 4360 is now fainter titan Hy, 

It is especially interesting that the lines X 4360 and X 5750 
should be the ones to change. The first measuiies of the spec- 
trum in August, 1892, showed unmistakably that it was the 
spectrum of a nebula. At first the lines X4360 and X5750 did 
not seem to exist in the old nebulae. However, photographs of 
their spectra showed the line X 4360 in five well-known nebulae ; 

* For the estimated intensities of the nineteen lines observed in 1892, see I*ub. 
^.51/14,245. 

49 



50 IV. W. CAMPBELL 

and careful visual observations showed the line X 5750 in three 
nebulae. These lines were strong in the Nava^ but relatively 
faint in the old nebulae. They have now become relatively faint 
in the Nova I 

The spectra of the well-known nebulae likewise have their 
anomalies. The lines X4472 and X4687 not only have very 
different intensities in different nebulae, but they seem to be 
entirely absent from some nebulae. The new star seems to be 
fast losing its anomalies ; its spectrum is not only nebular, but 
it is approaching the average type of nebular spectrum. 

The observed intensities of the line X4360, recorded in the 
above table, show that the decrease has been gradual rather than 
sudden. 

The lines still remain broad, as they always have been 
described. 

I have measured the wave-lengths of the two principal lines 
recently, with the following results : 

X894, September 7,]^" ^^58.7; t^=- 18- 
I X = 5006.4; » = -36*" 

1894, November 28, \ y= ^958-8; ^ = -"'" 
( X = 5006.8; V = —15*" 

These measures were made with reference to the iron line at 
X 4957.6 and the lead line at X 5005.6. With the micrometer 
wire in the positions of these lines it was seen with perfect ease 
that the star lines were less refrangible than the comparison 
lines. In August and September, 1892, the star lines were seen, 
on the contrary, to be much more refrangible than these same 
comparison lines. 

On the 1894, November 28, negative the wave-length of the 
Hy line is 4340.3, corresponding to a velocity of — 28*". Its 
wave-length on the 1892, September, negatives was 4335.8. The 
wave-length of Hy has therefore changed just as have those of 
the three principal lines. The wave-length of the line X4360 on 
the recent negative is 4364. On the 1892, September, nega- 
tives it was 4359. Thus this line has also shifted, as we would 
expect, along with the others. 



THE SPECTRUM OF NOVA AURIGjE 5 1 

As bearing upon any possible theory of Nova Auriga, perhaps 
it will not be out of place to say here what I said last winter in 
another journal/ The Harvard College Observatory has shown 
that both Nova Atiriga and Nova Norma at discovery possessed 
substantially identical spectra of bright and dark lines, similarly 
and equally displaced. Both diminished in brightness and both 
assumed the nebular type of spectrum. The new star of 1876 in 
Cygnus probably had nearly an identical history : passing from a 
bright star with a spectrum of bright and dark lines, to a faint 
object with a spectrum consisting of one bright line (undoubt- 
edly the nebular line X 5010, or the two nebular lines X 5010 and 
X4960 combined). We may say that only five "new stars" 
have been discovered since the application of the spectroscope 
to astronomical investigations, and that three of these have had 
substantially identical spectroscopic histories. This is a remark- 
able fact. We cannot say what the full significance of this fact 
is. One result, however, is very clear : the special theories pro- 
pounded by various spectroscopists to account for the phe- 
nomena observed in Nova Auriga must unquestionably give way 
to the more general theories. 

Mt. Hamilton, November 30, 1894. 

^Fub.A.S.P.t^Si, 103. 



THE MODERN SPECTROSCOPE. X. 

GENERAL CONSIDERATIONS RESPECTING THE DESIGN OF 
ASTRONOMICAL SPECTROSCOPES. 

By F. L. O. Wads WORTH. 

The special requirements which an astronomical spectroscope 
has to fulfil in addition to the usual ones possessed by the best 
laboratory instruments, are : 

1. The greatest possible degree of compactness consistent 
with a given degree of resolving power, which is usually deter- 
mined in advance by the character of the work which is to be 
done with the instrument. 

2. The highest possible efficiency as regards loss of light by 
reflection and absorption in the passage of the beam through the 
spectroscopic train, be it grating or prismatic. 

3. Lightness combined with an unusual degree of stiffness 
and rigidity between the various parts of the instrument and the 
telescope to which it is attached. Ease of attachment to and 
detachment from the latter is also of importance, particularly in 
the case of large instruments. 

To meet most satisfactorily these conflicting requirements 
in any particular case is a matter of some difficulty, and demands 
in the first place, careful consideration of the theoretical principles 
involved, some of which do not seem to be well understood by 
astronomers in general; and in the second place, no small degree 
of skill in optical and mechanical design. We will briefly con- 
sider these points in this, their relative order of importance. 

A. THEORETICAL CONSIDERATIONS.' 

In every case there must be a given condition as a start- 
ing point from which to work, and this will most natu- 
rally be a given degree of resolution, since upon this depends, 

' When this paper was written I did not know of the very excellent and complete 
paper by Professor Keeler {Sidereal Messenger^ November, 1 891), on the same subject. 

52 



THE MODERN SPECTROSCOPE S3 

not only the purity of the spectrum, but also, in the case of 
a bright line spectrum at least, its brightness.' The con- 
dition of a constant resolving power is a much fairer basis 
upon which to compare the efficiency of different spectroscopes 
than the condition, sometimes taken, of a constant dispersion, no 
matter whether the instrument is to be used for visual, photo- 
graphic or holographic observations. In the latter two cases 
a large linear magnitude of the spectrum is necessary, because 
of the finite size of the silver grains in the one case, and the 
finite width of the bolometer strip in the other. But this may 
always be secured under any conditions of resolution and dis- 
persion, by simply increasing the focal length of the observing 
telescope.' 

In what follows then, we shall suppose always (unless otherwise 
stated), a fixed resolving power, which we will call r. Let R, A , 
/% denote respectively the resolution, aperture (linear) and focal 
length of the telescope objective; / and tf, and /', a\ the cor- 

It will be found that some of the conclusions reached, particularly in regard to the nec- 
essary size of instrument, differ from those of Professor Keeler. This and other differ- 
ences are due in the main to the assumption of a different basis of comparison (constant 
dispersion as against constant resolution), and to the adoption of a different method 
of treatment (that of geometrical as against that of physical optics). 

' In discussing this point, Schuster, in his article " Spectroscopy '* (Ency, Brii, 
ss» 374)t says, " We come then to the conclusion, that for both narrow and wide 
slits the efficiency of a spectroscope depends exclusively on its resolving power." 
See also, Rayleigh, " Investigations in Optics with Special Reference to the Spectro- 
scope." Phii, Mag., 8, 261, 403, 407 ; 9, 40, 1879-80. 

* Linear magnitude of the spectral image must be carefully distinguished from 

dispersion, of which, as stated above, it is entirely independent. The dispersion is 

properly defined as the ratio between the change in the angle of deviation and the 

dB 
corresponding change in wave-length, vi%, by — ■-. It is, therefore, entirely independent 

a\ 

of the focal length of the observing telescope, and depends only upon the optical prop- 
erties and arrangement of the spectroscopic train. 

It is similarly equally independent of the resolving power R^ any value of the one 
being obtainable with any value of the other. With a given prism, for example, any 
▼aloe of the dispersion from a certain minimum (the dispersion at minimum deviation) 
to Qo may be obtained by simply changing the angle of incidence, while the resolving 
power is only slightly changed, 1. e. diminished, by the operation. It would seem that 
there ought to be no confusion as to the obvious distinctions between these three quan- 
tities, but they are nevertheless often confounded. 



54 F. L. O. WADSWORTH 

responding quantities for the collimator and observing tele- 
scope of the spectroscope ; 

^=^, and)8 = ~-. 

the angular apertures of the collimator and observing telescope ; 
w, the angular magnitude of the source of radiation; s, the width 
and h the height of the illuminated portion of the slit ; D, the 
dispersion of the spectroscopic train ; P, the purity of the spec- 
trum ; i\ its intensity, and b its intrinsic brightness. Further, 
in the case of the grating, let m denote the order of the spectrum 
observed; n, the number of lines in the grating of an aperture 
equal to a\ q, the grating interval; tf, the angle of diffraction, and i 
the angle of incidence. In the case of the prismatic train, ^ denotes 
the refracting angle of the prism, or prisms, if more than one is 
used ; 0, the angle of deviation of the refracted beam at minimum 
deviation; n, the index of refraction and N the number of prisms 
in the train. 

The most important consideration which will determine the 
resolving power r of any particular instrument will be the size 
(i. e., linear aperture) of the t-elescope with which it is to be used; 
for r should in general be proportional to R^ and this, for any 
particular part of the spectrum, is simply proportional to A. When 
the resolving power r is once determined by this general principle, 
modified in special cases by the conditions ot use, the focal length 
of the collimator of the spectroscope, which is the chief factor in 
determining the size of the latter, becomes determined by the 
relation >. F 

Since the resolving power r depends, in the case of a grating, 
only upon the total number of lines and the order of the spectrum, 
and in the case of a prism, only upon the difference between the 
thicknesses of refracting media traversed by the two extreme 
rays, it follows that we may obtain a given degree of resolution, 
either, i) by the use of a small aperture a with a very finely 
ruled grating, or a large number of prisms; or 2) by the use of a 
large aperture, with a more coarsely ruled grating, or a smaller 



THE MODERN SPECTROSCOPE 55 

number of prisms. Since the bulk, weight, and, roughly, the 
cost of the whole instrument is proportional to the cube of the 
aperture, while the rigidity obtainable with a given construction 
is inversely proportional to it, the advantage of keeping this as 
small as possible is readily appreciated. It becomes then a ques- 
tion of some importance to determine the relation between the 
aperture a and the dispersion, purity and brightness of the 
resultant spectrum. 

It may be readily shown that for any given value of r, the 
dispersion D is, in the case of both the grating and the prismatic 
spectroscope, inversely proportional to the aperture a. 

In the case of a grating, resolution is defined by the well-known relation 
r = mm, and since 0' (sin $-\-uai)^= mn\ we have at once 
-J av^ mn mn 

since a 'cos tf = a' = effective aperture of the observing telescope, and there- 
fore of the spectroscope. 

Hence, disregarding the small difference between a and a', consequent 
upon rotating the grating for different portions of the spectrum, we have 

a 
In the case of the prism, we have similarly, 

where /. and /« are the greatest and least thicknesses of refracting medium 
traversed by the extreme rays ; 



.•./,— 4 = ATfl 



aF 



«• sin* t 



and • ^ 

*"** sin - . 

.- • dn 

r — ^Na . ==r — 

-Ji— ««sin«^ 

Bat the dispersion n =z^=z^ ^ —— I 

* dX dm d\ dm \ 

Dn = N- 



dm 
d\ 



-J I— i»»sin' 



""^ d\ 



* * a as before. 



56 F. L, a WADSWORTH 

The purity P of the spectrum is independent of the aperture 
under the conditions of constant resolution. For purity is 
defined by the relation „ X 

and hence if ^ and r are constant, P is also constant. It fol- 
lows at once from this that for any given telescope (^=constant), 
and for any given degree of purity, /A? slit tuidik and therefore Ike 
total energy of the beam incident on the collimator is the same for all 
apertures; L e., for all focal lengths of the collimator. Hence, dis- 
regarding for a moment the losses due to transmission through 
the spectroscopic train, it also follows that the total energy in 
the spectral image of the slit, whether continuous or discon- 
tinuous, will be a constant, and will be equal to Ish, where 
/ is a factor depending on the intensity of the source of radia- 
tion, the aperture A and the focal length F of the large tele- 
scope. The intensity * of the image will be measured by the 
ratio between the area of the slit hs and the area of its image 
h' s\ and is therefore equal to 

. , hs 

€ being the factor of loss of energy by absorption, reflection, dif- 
fusion, etc., during transmission. In order to more completely 
define this ratio we need to express the value of A'j' in terms of 
hs and tp and P. It is evident that 

s^ = ts + £ + ff+G, 

where E is the effective broadening of the image due to the 
effect of diffraction ; //, the effective broadening due to the dis- 
persion ; and G that due to chromatic and spherical aberration, 
imperfection of optical surfaces and irregularities of dispersion 
produced by differences in density in the case of the prism train 
or by unequal spacing in the case of the grating. 

Similarly y = /\^£^G' ; 

.-. i=A-^. ^^. (0 



^s + £ + Jf+G)^li + £+G') 



THE MODERN SPECTROSCOPE $7 

The relative importance of the different terms in the denom- 
inator will depend upon the character of the spectrum under 
examination, the width of the slit s, and upon the quality of the 
instrument itself. In general in any good instrument G and C 
may be neglected in comparison with the others. Usually also 
the height k is large in comparison with the width s, and also in 
comparison with E ; and the expression ( i ) therefore reduces to 

in which the value of the terms E and //^must now be considered 
for particular cases. 

If the aperture is circular, the diffraction image of a ver- 
tical line (the bounding edge of the slit, for example) at 
the focus of the observing telescope, will be a band, the intensity 
of which at any point will be given by the expression 



/ = -clSJ where z = 



wax 



4 XV" ^ V ' 

X being the distance from the center of the image of the line 
and y, («) is Bessel's function of order unity. Tables of the 
values of y,*(«) r^ have been given by Lommel,* from which 
the curve. Fig. i, is plotted. The value of x for which /becomes 
zero is given by the roots of the equation y, {z)=o, the first of 
which gives x X 

Jf= 1.22 -/ =1-22-. 

a fl 

The total broadening of the slit image, due to diffraction 
from both edges, is therefore' 

2^=2.4^. 

'SCHLOMILCH, 15, 166,1870. Sec "Wave Theory," Rayleigh: Ency, Brit,, 
Hf 432. 

'The intensity of the first bright lateral band (see Fig. i) is only about ^ the 
intensity at the center, and it is therefore insensible to the eye or to the photographic 
plate under ordinary conditions. 

It must be noted that the curve, Fig. i, does not truly represent the falling away 
in intensity at the edges of the image of a slit of finite width, for in that case the dif- 
fraction figure is one produced by the superposition of the diffraction images of every 



58 



F. L. O. WADSWORTM 



But, owing to the rapid fading away of intensity at the edges 
of the image, the apparent broadening visually or photographi- 
cally will be somewhat less than this, not more than Jf = 2 ^ • 

The equivalent broadening may, so far as its effect on the 
intensity is concerned, be considered to be 

X^^E = ab = j. 

such that the area, ^^r// (Fig. i), is equal to the total shaded 




Flgl 

area under the curve, between the limits ji: = + «>; ji:= — 00, 

— 00 

For an aperture ^ =r ^ , which is not far from the usual value 
for an observing telescope, and ^ = y^y, we have 



7"^ = T = ^ 



.0075 or 0.0003 in. 



a quantity that can only be neglected when the slit-width is 
greater than -j-^ in. 

element of the slit. If, however, we disregard the effect of the first lateral band, it 
does truly represent the extreme broadening of the image at its edges, which is what 
we are concerned with here. 



THE MODERN SPECTROSCOPE 59 

In what precedes we have been considering the slit image as of uniform 
brightness across the whole width of the slit. There is one case of great im- 
portance in which this is not true, f . /., that of stellar images. Here the image 
of the star is so small that it falls entirely within the slit, and the effective 
width and height of the latter is simply the diameter of the star image. This 
will evidently be x» A ssm^-I-A", where «, as already defined, is the angu- 
lar magnitude of the star, and 

Since the intensity at the edge of the disk will be very small, we may for 
practical purposes write _ j. _ ^A^-^\ _ ^X 

'— — f ~ ^ ' 
In the case of large telescopes s is from 0*^.0 15 to o"".o2o.' 

The apparent (visible) broadening X' of the spectral image by the diffrac- 
tion of the observing telescope will in this case be considerably less than where 
the slit image is of uniform intensity across its entire width. It will, of 
course, vary with the absolute brightness of the star image, but will not in the 
majority of cases exceed one-half that found in the former case, or 

The effective broadening X^* of the spectral image is similarly less, and may 
be taken in this case as simply equal to ^ ' , since the intensity of the slit 
image itself varies according to the law /.=/(x), shown in Fig. i. 

Substituting these values in (i) (neglecting the terms G and G*\ we have 
therefore, in the case of stellar spectra, to a high degree of approximation. 



••=j"(-^)- 



* + 



|[t + ^] 



= ±y.(4)* L 



-?? 



(3) 



In considering the effect of H^ the broadening of the spectral 

image due to dispersion, two cases must be distinguished. First, 

that in which the value of H is small in comparison with usual 

values of j, which is the case of discontinuous or line-spectra ; 

second, that in which H is large, compared with j, the case of a 

continuous spectrum. 

' Star images generally appear larger than this because of the unsteadiness of 
Che air and consequent ''boiling** of the image. 



62 F. L. O. IVADSWORTH 

of the actual " width " of the lines, and the value of H is, there- 
fore, uncertain. The width of the reversed (bright) K line in 
the prominence spectrum has in certain cases been found to be 
about 0.2 tenth-meters. With a spectroscope of the resolving 
power and angular aperture just assumed, the value of Z will 
therefore be 

Z= 25000 X 15 X .2 X 10-^ = .0075 

.'. H= o"".oo37. 
In most cases the resolving power of an astronomical spectro- 
scope will be considerably lower than this, so that even if AX is in 
particular cases much larger, the resulting value of H will in gen- 
eral be small compared with the slit-widths which it is generally 
necessary to use. An important exception to this is found in the 
case of the spectra of nebulae, which will soon be considered. 

We have already found that the value of E is about .0075 
(for )9= 1^), or in the case of star spectra about \ this amount, 
(0™".004). In most cases these quantities may be neglected in 
comparison with j, and the expression for i becomes simply 

.•=M|)". 

Under conditions of constant magnification (^) becomes constant, 

and the brightness b of the spectrum, as viewed by the eye, is 
therefore nearly constant for all apertures and for all slit-widths 
greater than o"'^.025 = 0.00 1 in. For widths less than this it is 
necessary to consider also the efiEect of the terms H and E. 

It is evident that the maximum intensity at the central portion 
of the image will be reached when the width s is equal to 



!(?+^) 



But (see preceding footnote) 

2k r 

X= -g-, and (see above) ^=o AX ; 

for maximum intensity (or brightness) 



^ = J + ^^- (4) 



THE MODERN SPECTROSCOPE 63 

Substituting this value of s in the expression for the purity, we 
have for the condition of maximum brightness and maximum 
purity ^ X 



2XXrAX ' ' 

which shows that the purity of the spectrum can never, even with 
an absolutely monochromatic source, exceed 50 per cent, of the 
resolving power, unless the brightness be sacrificed by making 
the angular width of the slit less than the angle subtended 
by a wave-length of light at a distance equal to the aperture a 
of the collimator.' 

To obtain an idea of the relative importance of the term rAX 
as compared with X, let us consider the case of a solar promi- 
nence line observed, as before, with a spectroscope of resolving 
power r « 25000. We have then 

rAX » 25000 X 2 X io~* = .0005 =s X ; 
or, in this case, the width of the slit which gives maximum illu- 
mination is twice as great as that given by the limit x»-. 

With still higher resTolving powers, or larger values of AX, the 
importance of the second term with respect to the first becomes 
even greater than this. In the case of the nebulae, for example, 
it is found that the brightness of the lines of the spectrum con- 
tinues to increase until the slit- width is from i^ to 3 or 4 times 
according to the resolving power employed) as great as what 
has been (erroneously) regarded as the theoretical slit-width for 

' It it vinally stated that the width of the slit for maximum illomination is 
is^, and that the increased brightness actually observed when the slit is 

widened^ beyond this point is only apparent, and due to physiological causes. 
Schuster (Eney. Brii,,n^ 374) says: '*The maximum illumination for any line is 
obtained when the angular width of the slit is equal to the angle subtended by one 
waTC-length at a distance equal to the collimator aperture. In that case ^=X . . . 
If the Tiiual impression depended only on the intensity of illumination, a further 
widening of the slit should not increase the yisibility of a line. As a matter of fact, 
spectroscopists generally work with slits wider than that which theoretically giyes full 
illumination. The explanation of the fact is ph3rsiological, yisibility depending on the 
apparent width of the object'* It is evident, howeyer, from the preceding, that ' =^ 
is not the tine limit of slit width for maximum illumination, but there is an actual 
increase of brii^tness up to the point / = t'(X -f- rAX). 



64 F, L, O. IVADSWORTH 

maximum intensity. If observations of the relative intensity of 
the central portion of the spectral image with different widths of 
slit were made with sufficient accuracy, with a bolometer or 
photometer, for example, we could determine from the above 
relation the value of AX, and hence obtain some knowledge of 
the conditions of temperature and pressure in these distant 
masses of gas.' For if we call ^o the value of s for which the 
intensity becomes constant, we have from (4) 

r 

The only observations on s^ which are immediately available are those o£ 
Keeler on the Orion Nebula." Professor Keeler says : " Experiment showed, 
however, that with the same exposure the density of the photographed lines 
began to fall off sensibly when the slit- width was reduced below .001 in., or when 
it was still three times the theoretical width." (The theoretical width here 

referred to is that given by Schuster) The discrepancy is, I suppose, 

to be attributed to the spreading of the photographic image; .... the 
physiological effect which makes a line appear brighter when the slit is 
widened in visual observations seems to be analogous to this photographic 
action. The slit-width which gave the best results vas about 0.0015 in." 

If we take Jo to be o.ooi in. = ""'.025 (allowing 0.0005, or i, of the increased 
effect as due to photographic irradiation), we have for AX 

0.025^^.0005 . 

AX.= j: 

In Professor Keeler's instrument with which these observations were made, 
^ = ,1^, and r (a single 60° prism of white flint, with an aperture of i . 1 3 in., was 
used) was about 5000. ' We have, therefore, AX, = 2.6 tenth-meters, or 
nearly four times the width of the red hydrogen line at a pressure of xoo*"". 

If the preceding data are trustworthy, this result indicates either that the 
pressure is much greater than is ordinarily supposed, or that the whole mass 

' A much better method is, of course, to examine these radiations by means of 
a small wave comparer, attached to the observing telescope in place of the usual eye- 
piece or photographic plate. A preliminary trial of this method has already been 
made by Michelson and Hale on the bright lines of a solar prominence with very 
promising results, although the particular apparatus used was ill adapted to the 
purpose, and further experiments have therefore been postponed until a more suitable 
instrument can be constructed. 

• "On the Spectra of the Orion Nebula and the Orion Stars," J. E. Keeler, A, 
and A,^ October, 1893. 

s Calculated from data given by Professor Keeler, in a paper describing the 
Allegheny spectroscope. — A, and A. 



THE MODERN SPECTROSCOPE 65 

of the gas in the nebula is in a state of violent agitation (which would broaden 
the lines by reason of varying velocities in the line of sight). Or it might 
indicate that the lines themselves, which appear to be single, are in reality 
made up of a number of components. Further theorizing on the question 
is useless until more exact data are obtained by one of the methods 
previously indicated. 

2) Case of continuous spectra. If AX be large in compari- 
son with X, I. e,, if the source of radiation sends out waves 
whose length varies regularly and continuously from X, to X^, 
the value of H becomes so large that in comparison with it all 
the other terms in the denominator may be neglected. In this 
case, if we suppose the intensity uniform from end to end of the 
spectrum, jy=Z = ;(X.-A.) . 

We then obtain from (2) 

which shows that in this case the intensity is proportional to the 
width of the slit s, and inversely proportional to the resolution. 
But it also shows that for a given purity P (j^= constant), and 
given resolution r, the intensity, just as in the case of the dis- 
continuous spectrum, is independent of the aperture a. Hence 
the brightness under constant magnification will also be inde- 
pendent of the aperture and hence of the dispersion of the 
spectroscope. 

The only difference, therefore, between this and the preced- 
ing case is that in the first the intensity and brightness is inde- 
pendent of the slit-width (above a certain minimum width, 
whose value has been already discussed), and in the second, both 
of these quantities are directly proportional to the slit-width. 

In the comparison of spectroscopes of different focal lengths 
but of constant resolving power, it only remains to determine 
the relative losses during transmission of the beam through the 
spectroscopic train, so far as these are affected by a change of 
aperture. As regards these losses we have unfortunately very 

' In the case of star spectra the formulae are the same with the exception of the 
fwtor I [see (3)]. 



66 F. L. a IVADSWORTH 

little data in the case of either the prismatic or the grating 
train. In the former case the loss during transmission is 
made up of two factors : one, the loss by absorption in the 
refracting media, the other, the loss by reflection from the sur- 
faces of the prisms. The loss by absorption may easily be 
shown to be the same for all apertures, since the thickness of 
refracting media traversed by the different rays is in each case 
the same.' The relation between the index of refraction and 
the absorption, either total or local, has been experimentally 
determined only in the case of very few kinds of glass, and even 
for those, not with a sufficient degree of completeness. It is to 
be hoped that experiments will soon be made for the sake of 
more completely determining this relation. The second ele- 
ment of loss by reflection increases as the aperture diminishes ; 
for, as we have already seen, the number of prisms N^ and 
hence the number of reflecting surfaces* varies inversely with 
the aperture a. This loss, however, increases much less rap- 
idly than the number of prisms, because after the first few 
reflections the light becomes almost completely polarized at 
right angles to the plane of incidence, and is therefore almost 
completely transmitted by the succeeding surfaces. For the 
sake of a more complete representation of this important fact, 
the following brief table, abridged from tables which I have 
recently prepared, is presented. It gives the per cent, loss 
of light by reflection in the case of a prism train consisting of 
iV^ prisms of flint glass (iV],= i.6) of 64^ angle, on the assump- 
tions: i) that the incident light may be regarded as made up of 

'Pickering {Am, Jour, ^ 14, 1868) makes the following statement, which has 
been quoted in Scheiner's Astr^momicai Spectroscopy (Frost's translation): 

" In spectroscopes of the same dispertian and from the same glass the loss of light 
by absorption is the same." lliis is eyidently only true for the special case in which the 
two spectroscopes have the same aperture. The more general principle is that giyen 
above ; vfs., that the absorption is die same (for any given kind of glass) in all spectro- 
scopes having the tame resolving power^ no matter what the divpersion or the aperture. 

*The number of reflecting surfaces =2A^, in case the train is a single transmis- 
sion one of simple prisms, and =4^A^4~2) if the train is a multiple transmission 
one of the Littrow or modified Littrow type, and is made up of compound prisms, Q 
being the number of times the beam passes through the train. 



THE MODERN SPECTROSCOPE 



67 



two beams of equal intensity polarized at right angles to one 
another : 2) that the loss by reflection may be correctly repre- 
sented by Fresnel's formula, and 3) that the angle of incidence 
is the angle of complete polarization, or so near it that the loss 
by reflection of that portion of the beam polarized at right 
angles to the plane of incidence may be disregarded. The 
total loss can therefore never exceed 50 per cent. In the case 
of prisms of white flint (11=1.6) this last condition will be satis- 
fied if the refracting angle of the prism is 64**, and very nearly 
satisfied if the refracting angle is" 60^ (see Table). 

Table II. 

Loss of Light by Reflection from the Surfaces of a Train of 

Prisms of 64'' Refracting Angle and Index 1.6. 



Number of 


Number of 


Lossby 


Percent, of 


Loss by Reaection 


Prisms 


SnrfMBS 




TomlLoss 


for Prisms of 60" X 


I 


2 


.174 


35 


.147 


2 


4 


.287 


57 


.252 


3 


6 


.361 


7* 


.328 


4 


8 


.409 


82 


.382 


5 


10 


.441 


88 


.422 


10 


20 


.493 


98 


.504 


00 


00 


.500 


100 





It appears from the above table that more than \ the total 
loss of light is caused by the first prism, and that very little 
additional loss is caused by the addition of any number of 
prisms after the third, provided only that condition 3) is satisfied. 

As regards the loss of light in the case of a grating, we have 
still less data, either theoretical or experimental, to guide us. 
The loss by diffusion will in general tend to increase with the 
closeness of the ruling, but it may easily happen that this may 
be more than counterbalanced by the more perfect concentration 
of the light in one particular spectrum.' Irregularities of spac- 

■From Pickering's Tables. Amer.Jomr.t 451 1868. 

'The concentration of the light in any particular spectrum is dependent onlj 
upon the form of the cross section of the rulings. This is not at present, nor is it 
likely soon to be, a determinate problem constructively ; although gratings are often 
obtained, either accidentally or by trial settings of the ruling point, which show a high 
degree of efficiency in this respect. 



68 F. L. O. WADSWORTH 

ing, however, have a much more injurious effect upon definition, 
and therefore upon purity, in finely than in coarsely ruled gratings ; 
and for this reason it is very difficult to produce a satisfactory 
grating with more than 1500 lines to the mm. (y = o"".ooo67). 
So far we have considered the case of a constant angular 

aperture, ^ = ^i of the telescope, and a constant degree of 

resolution r. Let us see what will be the effect of varying either 
or both of these quantities. 

I. Effect of varying ^. It is evident that the effect of 
changing ^ is simply to change the value of /, the intensity of 
the image on the slit. In considering this effect it is necessary 
to distinguish between those cases in which oi, the angular mag- 
nitude of the source, is insensible, and those in which it has a 
considerable value. In the first case the energy-gathering power, 
and therefore the total energy in the image, will increase in the 
ratio ^', while the area of the image will increase in the ratio 

The mean intensity of the image / will therefore vary as 

/ ^ V^' or I = k-- — ^•, 

V»^+ 2X^ ^ M+ 2X)- ^ 

where k is an absolute constant depending for its value only on 

the intensity of the source of radiation. The effective width of 

the slit is in this case simply the diameter of the image, which 

is, as we have just seen, w^ -f 2X 

The purity of the spectrum is therefore constant, for 
« X X „ 



and faA is always small compared to 3X. 

Under the condition of constant purity we have, therefore, 
for the intensity from (3) and (5) : 

a) For continuous spectra 

f=-/c(y ,, ^,, ^->^€/9M'- ,, ' ,, ; (6) 

3 ^r K^.— ^) 3 2rX(X,— X.) ^ ' 



THE MODERN SPECTROSCOPE 69 



*) For discontinuous spectra 



'-Ai) 



= --fc,JM'- 



?H_ -2 ' 2X(2X + irAA)" (7) 

'■^> 

That is, the intensity is in all cases independent of the angular 
aperture ^, but increases as the square of the linear aperture A of 
the telescope. The necessity for a large aperture in order to 
obtain brilliant stellar spectra is at once evident. In order, how- 
ever, to obtain pure spectra a large resolving power is absolutely 
necessary, for, as we have just seen, 

P ^, 

and the purity of the spectrum can therefore never exceed one- 
third of the theoretical resolving power. This explains to some 
degree the want of sharpness always observed in stellar spectra, 
even when viewed under the most favorable conditions. 

If we follow the rule proposed in the first of this article, and 
make the resolving power of the spectroscope proportional to the 
aperture of the telescope, then r^=r^ A^ where r^ is the resolving 
power for unit aperture. Under these circumstances we have for 1, 

2 A 

f=- ^^ ' wv YY" ^^^ continuous spectra; 

1 A 

/ = - ktj^' r for discontinuous spectra; 

2 ./2X , I ..\ ^ 

which shows that under these conditions the intensity increases in 
a somewhat smaller ratio than the aperture (because of the increase 
of c with r, on account of the additional losses, due to absorption, 
reflection and diffusion, which attend the use of a larger aperture 
and a higher resolving power). 

Equations (6) and (7) show that the intensity (and therefore 
brightness) of the spectrum varies inversely as the wave-length. 
In the case of prismatic spectra this is compensated by the 
fact that the resolving power of a prism varies inversely as the 
cube of the wave-length. In the case of the continuous spectrum, 
equation (6) becomes, therefore. 



70 F. L. O. IVADSIVORTH 






where r^ is the resolving power of the instrument for some 
particular part of the spectrum. In addition to its higher 
efficiency as regards the concentration of light in only one 
spectrum, the prism train has therefore this other important 
advantage over the grating for the purposes of photog^phy, 
vvs,^ the more uniform distribution of actinic intensity in the 
spectral image. 

The preceding considerations indicate the correct lines for 
the construction of star spectroscopes, which strangely enough 
do not seem to have been generally understood, or at least 
regarded in the design of existing instruments, perhaps because 
it has always been necessary to use them on existing telescopes. 
The importance of the work, however, would now certainly seem 
to warrant the use of specially constructed image lenses or con- 
densers instead of the usual telescope objectives, which are usually 
considerably more accurately corrected than they need be for 
this work, and could therefore be more advantageously used for 
other purposes. The image-forming lens or "condenser" of a 
star spectroscope should have : i) The largest possible linear 
aperture A, in order : «) to increase the intensity of the star image ; 
^) to increase the value of the term laA in comparison with X, 
and thus secure a more uniform distribution of intensity across 
the breadth of the image. 2) A very large angular aperture 
^, I. e., a very short focal length F, in order to make the width of 
the image / «^ + aX ^ 

and therefore the effective width of the slit, as small as possible, 
and also in order to make the whole arrangement as compact as 
possible. These two requirements are best met by the use of a 
parabolic mirror, which has the further advantage of having no 
chromatic aberration and requiring no change of focus for differ- 
ent wave-lengths. 

It does not seem at all impossible to make such reflectors (silver 
on glass) with a linear aperture of 6, 8, or even 10 feet, with a focal 



B 



PLATE V 



•If- 

-hi 



;j 






,.:r^ 



^i^tSf:::.:—=-.... 



THE MODERN SPECTROSCOPE 7 1 

length only three to four times the aperture.* This would 
require the use of a concave (parabolic) mirror in place of a lens 
for the collimator of the spectroscope, but as I have already 
pointed out,' this is attended by no disadvantages ; on the con- 
trary, it has certain advantages, even when the image is formed 
out of the axis of the mirror. 

In Plate V,-^, B, C,are shown diagrammatically some forms of 
stellar spectroscopes designed on the above lines. If it is con- 
sidered desirable to use a lens instead of a mirror for the collimator, 
this may be done by the introduction of a convex mirror (or con- 
cave lens) in the cone of rays before they reach the focus (as in 
Plate V, C ), although this can only be done at a sacrifice of the 
sharpness and brightness of the star image. On this ground 
form^ A and B are certainly to be considered preferable. 

If the angular value oi of the source be so large that the width 
of the image s* be greater than the width of the slit s, fnA 
becomes large compared with X, and the ratio 

(— J-7-— r-)' becomes simply (-)' , and therefore 



^A + 2X 






In this case the purity of the spectra will vary with ^ in the 
ratio X 

It may, therefore, be maintained constant in either one of two 
ways: i) By diminishing the width of the slit s, as the angular 
aperture ip increases, so that the product j^ remains constant. 

' The large reflecting telescope made by Mr. Brashear for the Smithsonian Astro- 
phjrsical Observatory has an aperture of 50^ and a focal length of only i meter, 
I. e., the ratio of focal length to aperture is only 2 to I. This instrument gives very 
satisfactory definition, better in fact than would be necessary for its use as a 
condenser. 

•"An Improved Form of Littrow Spectroscope;" F. L. O. Wadsworth, PAi/. Mag. 
July, 1894. I have since learned that this form of spectroscope had been previously 
suggested by Lippich and Ebert in the Ziit, fur Inst. It is also now being used by 
Messrs. Kayser and Runge (see Pulfrich, Zeit, fur InsL^ October, 1894). I regret 
that I did not know of the papers of the gentlemen first mentioned at the time of 
writing the article, so that I might have accorded them the claim of priority. 



^2 F. L. O. ]VADS]VORTH 

2) By increasing the resolution, as ^ increases in the proportion 
x^ 4* A, J remaining constant. 

By the first method we diminish the intensity (of a continu- 
ous spectrum) in the ratio -7 by closing the slit ; by the second 

we diminish it in a somewhat smaller ratio: 

I 
i^ + A' 

by an increase in r, but the gain in the last case is counter- 
balanced by the increased loss due to greater absorption or 
diffusion, so that the final result is about the same in either 
case. 

We have, therefore, under the condition of constant purity, 
as before : 

a) For continuous spectra, 

_ ( J == j^^ for constant resolution, 

^^^^^^\r = r^{s^-\-\) = r^{si^) for constant slit-width. 

••• •■ = '*(|)";(»r^='^K-X^ ,f.r.=co,»t. 

b) For discontinuous spectra, 

b^ For wide slits (greater than o"".02) 

^,) For narrow slits (less than 0™*.02) 
I = /. {t\ £ = itl)' 1 ; (10) 

which shows that in this case the intensity is independent both 
of the angular aperture ^ and the linear aperture A of the tele- 
scope, and, with a given observing telescope and a given spectro- 
scope, depends, as before, only upon the value of k^ that is, upon 
the intensity of the source of radiation. 

Equation (10) shows that in the case of bright line spectra 
from sources having a considerable angular magnitude (nebulae, 



(8) 



THE MODERN SPECTROSCOPE 73 

comets, solar prominences, faculae, etc.) the intensity and, there- 
fore, brightness will necessarily diminish as the resolving power 
increases. Pure spectra can then only be obtained when the 
source is very bright, and a small telescope is just as advanta- 
geous as a large one in viewing the spectra of such objects. 

Uu of the teieseope objective as a {so-cailed) "condensing"* tens, — Since 
the intensity of the spectrum of a source having a finite angular magnitude is 
independent both of the angular aperture f and the linear aperture A of the 
telescope, it is evident that the term *' condenser" is in this case inap- 
propriate. Indeed, we may reduce the latter to infinitesimal dimensions, or, 
what practically amounts to the same thing, dispeAse with it altogether 
without loss of light, provided only: i. That the angular magnitude «# of the 
source is equal to the angular aperture f of the collimator ; 2. That the 
source is of uniform brightness ; 3. That the form of the source is geometric- 
ally similar to the aperture of the spectroscope. 

The resulting spectrum is necessarily the resultant or integrated effect 
from all parts of the surface, and it will, therefore, be more or less impure, 
according as the radiation conditions of different portions of the source differ 
from each other. Thus, in a solar spectrum formed by such a spectroscope, 
the dark lines will all be broadened by an amount equal to 

^ AXv , where AXv 

is the difference in the wave-lengths of a given radiation from the east and 
west limbs of the Sun, produced by reason of the rotation of the latter with 
velocity v. If we assume the equatorial velocity to be about 2^ per second, 
AAv is about .08 tenth-meters, or an amount equal to about J the 
breadth of one of the sodium lines at atmospheric pressure. If the 
highest degree of purity is required, the use of an image lens or 
condenser is, therefore, indispensable. But it may also be shown that, in the 
case of the Sun or other heavenly body surrounded by an absorbing atmos- 
phere, the use of a condensing lens is necessary, in order to secure the maxi- 
mum intensity of spectrum. This may be shown as follows : 

Let e denote the quantity of energy received normally per unit area at 
the surface of the Earth from unit area of the Sun. Then the total energy 
received per unit area from all parts of the Sun*s disk will be, since the 
angular magnitude of the radiating surface is small, 

or, if we suppose the falling off in intensity to be symmetrical about the 
center, more simply .R 

R being the radius of the solar disk. 



74 F. L, O. iVADSIVORTH 

If no condenser is used, and the spectroscope train has an aperture of 
sufficient size to receive all the light which passes through the slit, the total 
energy incident on the collimator will evidently be 

vxdx . 






will, in this case, be 

twA' 



If a condenser is used, the energy transmitted to the image on the slit 
from each element of the solar surface will be ^==^cri4V, and if, again, 
the spectroscope has an angular aperture large enough to transmit all the 
light received from the slit,' the total energy incident on the collimator 

where jt,, jr„ and y„ y,, are, respectively, 

- ^ t - ^ t 

^ y ^ y 

(tt (a> and Ti* r»t being, respectively, the coordinates of the horizontal and 
vertical edges of the slit referred to the center of the solar image as origin, and 
D, the distance of the Sun. If the solar image is placed symmetrically on 
the slit, and its diameter is not greater than the length of slit opening, we 
have simply /> s' 

Further, the width of the strip the coordinates of whose edges on the solar 
surface are x„ ;r„ is so small that in the integration for x, 0- may be consid- 
ered constant. A first integration gives us then 






^s' Cffdy = iarA'^s'jffify, 



and therefore 



sA j aytfy 



' If the diameter of the solar image on the slit plate is small, compared with the 
focal length of the collimator, we must have, in order to fulfil this condition, a ' =/f + A. 
Hence the angular aperture (vertical) of the spectroscope must be 

while the horizontal aperture will be as before, 

^ / /• 



THE MODERN SPECTROSCOPE 75 

To determine the value of this ratio in any particular case we must know 
the law of variation of 9 or the form of the function j- = ^(y). In the case of 
the Sun, the experiments of Vogel ' show that for the brightest part of the 
spectrum, the curve of j^=0^ is very nearly a circle having its center coinci- 
dent with the center of the disk.' 

Hence, if 9^ denote the intensity of radiation at the center, the intensity 9 
at any part of the Sun*s disk in terms of the distance y from the center will be 
very nearly /-- 

Substituting this value for 9, and integrating, we obtain for the two pre- 
ceding integrals » _ 



^r^ 



fv 



9ydy = ~T^ 



Substituting these values in (9), we obtain 



/^ " 16 M R % Sia h ' 

But for constant purity of spectrum in the two cases we must have equiv- 
alent slit- widths, or , w / 

and we have simply for constant purity 

P'~~4 h * 
that is, for an aperture of the condenser equal to the height of the slit used 
without a condenser, the intensity of the spectral image will (neglecting the 
absorption of the glass of the condenser) be about 25 per cent, brighter with the 
condenser than without. If we are concerned not with the brightness of the 
spectrum but with the total intensity in the image of any one line, as in 
bolometric work, for example, the immense advantage of the use of the con- 
denser is evident whether the radiating surface be of uniform brightness or 
not.' For it is possible to make the aperture A at least ten times as great as 
the greatest possible length of slit which can be used without a condenser, 
(on account of the necessary limitations in size of prisms and focal length of 
collimator). We can, therefore, by the use of the latter, obtain under the 

> These experiments were confiimed in a striking manner and by an entirely 
different method by Michelson in bis measurements of the angular diameter of the Sun's 
linage by interference methods. — See '* Application of Interference Methods to Astro- 
nomical Measurements," A. A. Michelson, PAi/, Afag., July, 1890. 

* Figures 4 and 5, /h'd, 

3 If 0- is constant, the ratio ^ reduces to simply c^, which shows that the bright- 
ness of the spectral image is the same with a condenser as without (or slightly less, on 
account of the absorption of the additional lens). 



7^ /^. L. O, WADSWORTH 

same conditions of resolution, purity and longitudinal magnification, a spectral 
image whose height,' and therefore total intensity, is from ten to twelve times 
greater than could possibly be obtained without it. To increase the intensity 
per unit width to any required degree, it is only necessary to contract the 
spectrum vertically by using a cylindric lens of proper curvature, placed with 
its axis at right angles to the length of the slit, either in front of the latter or 
between the objective of the view telescope and its focal plane. The former 
position is always preferable when the best definition is desired (as in photo- 
graphic work), although it increases the necessary vertical aperture of the 
spectroscope by an amount inversely proportional to the distance of the 
cylindric lens from the slit. 

The use of the condenser has the further advantage that with it much 
smaller focal lengths may be used, and the whole instrument thereby made 
more compact and manageable. A practical illustration will perhaps bring 
out these points more clearly. The largest optical prism which has yet been 
made, so far as I am aware, is one recently furnished by Brashear for the 
Smithsonian Astrophysical Observatory. It is made of rock-salt from a block 
exhibited by the Russian government at the World's Fair, and has a clear horizon- 
tal aperture of about 9^^ and a vertical aperture of 19^. When no condenser 
is employed, the greatest height of slit which can be used with this prism 
without loss of light is a little less than lo"^. The necessary focal length of 

the collimator is — =11", or about 35 feet. The height of the spectral 

image with an observing telescope of say 2" focal length will be about i}^. 
Now if instead of this arrangement we use a condensing lens of say 50^ 
aperture and 4"* focal length, the height of the slit, in this case the 
diameter of the solar image, will be reduced to about 3^^^. For the same 
horizontal aperture of prism as before, the focal length of the collimator will 
be only \^ and its vertical aperture will be about 13*", instead of 19*=" as 
before. The height of the spectral image, for the same observing telescope 
as before, will be about 9"", and the total energy will be over six times as 
great as before. 

These conclusions have all been verified experimentally both in the visible 
and in the infra-red spectrum by means of a bolometer placed at the focus of 
the observing telescope. 

The experiments were made with condensing lenses of different apertures 
and different focal lengths under the conditions of constant resolution and 
constant purity, 1. e,, with the same prism, and with a slit-width inversely pro- 
portional to the angular aperture of the condenser, and hence directly propor- 
tional to the focal length of the collimator (the horizontal aperture of the 

« We have >fc'=>l ^ = —A or tor a given aperture of the observing telescope, 
the height of the spectrum is directly proportional to the aperture of the condenser. 



THE MODERN SPECTROSCOPE 77 

latter remaining constant). The same observing telescope was used in each 
case, and a cylindric lens with its axis at right angles to the refracting edge of 
the prism was placed just in front of the focal plane, in order to reduce the 
height of the spectrum in each case to the length of the bolometer strip 
(about 1*"). 

B. MECHANICAL CONSIDERATIONS. 

No discussion is necessary to show that all mechanical con- 
siderations are favorable to a reduction in the size of the instru- 
ment to the lowest possible limit consistent with its efficient 
performance. From our preceding discussion of the theory of 
the spectroscope under the condition of constant resolving 
power, we have seen that the only decrease in efficiency which 
accompanies a decrease in size is that due to a slight additional 
loss of light (by increased diffusion, due to a finer ruling in the 
case of the grating, and by an increased number of reflections 
in the case of the prism train), and a possible impairment of 
definition, in the case of the grating, by the increased effect 
of the unavoidable errors of spacing ' when the grating space 
becomes very small. The limit to reduction in size is in all 
cases determined not by this reduction in efficiency, which is 
comparatively slight, but by mechanical and optical difficulties 
of construction. In the case of the grating, for example, il 
is practically impossible to rule a satisfactory grating with more 
than 40,000 lines to the inch, and even gratings with 30,000 
lines are not commercially obtainable. 

If we require, then, a resolving power of 25,000 in the first 
spectrum, we must use in the case of a grating spectroscope 
an aperture of at least |- inch. Granting, however, that a good 
grating of 40,000 lines per inch is available, a spectroscope of 
this size would be in every way as powerful and efficient as one 
with an aperture of 2^ inches, using a grating of the ordinary 
fineness (10,000 lines per inch), and far more rigid and con- 
venient to use, while the weight, bulk and (approximately at 

' Even if the screw of the dividing engine were absolutely perfect, errors of 
spacing would still occur, due to inaccuracies in the slides which carry the ruling 
point, to changes of temperature, to differences in hardness of different portions of 
the ruled surface, to slight errors in the surface, and to many other causes. 



78 F. L. O. WADSWORTH 

least) the cost of the smaller instrument will be only -^-^ that of 
the larger one. 

The same considerations hold with respect to the prismatic 
form. Here the number of prisms (of any particular kind of 
glass, and of any given refracting angle) requisite to secure a 
given degree of resolution varies inversely as the aperture. 
The reduction in aperture is, then, in this case only limited by 
the number of prisms which it is possible to employ. With 
prisms of dense flint of 6o° angle, the resolving power is about 
4000 per inch of aperture for a single prism. To obtain a 
resolving power of 25,000 with an aperture of 2\ inches 
would, therefore, require three prisms. To obtain the same 
degree of resolution with an aperture one-quarter as large, 
viz. \ inch, would require, therefore, twelve prisms. The 
volume of glass required would, however, be only one-six- 
teenth as great, and the total area of optical surface one-quarter 
as great. 

The definition would be likely to be better the smaller the 
instrument, because the glass would be more homogeneous in 
small than in large blocks, the small surfaces could be more 
accurately worked, and with a large number of surfaces the 
errors would tend to compensate each other. The increase in 
the loss of light by reflection, in using twelve rather than three 
prisms, would be only 14 per cent, of the incident light (see 
Table I), while the loss by absorption would be the same. 
The principal mechanical difficulty is in mounting the prisms 
and automatically maintaining them at minimum deviation, with 
a sufficient degree of accuracy for the purposes of mechanical 
measurement. Usually, however, this is not necessary, accurate 
measurements being usually made by means of a standard com- 
parison spectrum, which is observed at the same time, or photo- 
graphed on the same plate. 

The limit to reduction in size of aperture is, then, in this 
case even lower than in the case of the grating spectroscope. 
For by adopting the multiple transmission form of spectroscope 
(the original Littrow, or better. Young and Lockyer's modifica- 



THE MODERN SPECTROSCOPE 79 

tion of it),* there is no difBculty in using the equivalent of 

twenty or thirty prisms, and thereby obtaining a resolving power 

of from 50,000 to 60,000 with a \ inch aperture, or from 30,000 

to 35,000 with only a \ inch aperture. There is in this case, 

however, considerable additional loss by the repeated reflection 

of the beam by the reflecting mirrors, or right-angled prisms, at 

the ends of the train. 

University of Chicago, 
November, 1894. 

' See Schellen, Spectraianalyse, i, 231. 

A very compact form of an instrument of this type has been made by Gnibb 
(M. AT., 31, 36.) 

These forms and others have been discussed by the author in the paper already 
referred to. {PAil, Mag,, July, 1894.) 



Minor Contributions and Notes. 



THE ASTROPHYSICAL JOURNAL. 

In a paper bearing the above title, published in the first number 
of Astronomy and Astro-Physics (January, 1892), the reasons which 
had prompted the publication of a journal of astronomical physics 
were enumerated, and evidence was adduced to show that considerable 
support might be expected for such a venture. It had been my inten- 
tion to establish a separate astrophysical journal, but the uncertainty 
of such an undertaking led to an acceptance of Professor Payne's pro- 
posal of a union with the Sidereal Messenger^ and Astronomy and 
Astro-Physics was the result. The contents of the thirty numbers 
published during the three years which have elapsed since that time 
offer sufficient testimony to the usefulness of the composite journal. 
From the outset the editorial supervision of the departments of Genercl 
Astronomy and Astro-Physics was kept entirely distinct. The policy of 
the latter department was determined by myself and my associates, 
Professors Keeler, Crew and Ames, while the selection of all other 
matter published in the journal was made by Professor Payne and 
those who were associated with him. No attempt was made to draw a 
hard and fast line between the two departments. Had this been done, 
and a strict definition of "astrophysics" adhered to, a large part of 
the matter published undef General Astronomy would have appeared 
in the other department of the journal. It was thought best, however, 
to confine the scope of Astro-Physics to the more technical subjects 
connected for the most part with spectroscopic work. 

In returning to the original idea of a purely astrophysical journal 
we are simply following out a long-cherished plan. Few who appre- 
ciate the true scope of astrophysics, and have its best interests at heart, 
will deny the advisability of devoting an entire journal to this, the 
most fascinating and at the same time the most rapidly advancing 
department of astronomical research. In spite of the existence of 
physical and astronomical journals of the highest class, the astro- 
physicist or spectroscopist is at a loss to know where to publish in 
order to reach the audience he desires. Should he choose an astro- 

80 



MINOR CONTRIBUTIONS AND NOTES 8 1 

nomical journal, he will find that his paper will remain unread and 
unknown by a very large majority of physicists — the very men who 
are, perhaps, best (;pmpetent to appreciate its true value. A purely 
physical journal is not less evidently an unsuitable place for papers 
treating of solar or stellar investigations, even though these investiga- 
tions be prosecuted by the methods of the physical laboratory. 
Recent papers on the radiation of gases and the validity of Kirchhoff's 
law have all appeared in physical journals, but the subject is one 
not less interesting to the astronomer than to the physicist. The 
same might be said of scores of other papers dealing with spectroscopic, 
bolometric, radiometric, photographic and photometric researches 
conducted in the laboratory, but finding their most important applica- 
tions in the elucidation of astronomical phenomena. The astronomer 
and physicist should be able to meet on common ground, and this 
only an astrophysical journal can supply. 

During a recent visit to many of the observatories and spectro- 
scopic laboratories of Europe the writer enjoyed an excellent oppor- 
tunity to discuss with both astronomers and physicists the plan of 
founding such a journal. At Potsdam Professor H. C. Vogel, Director 
of the Astrophysical Observatory, and Professors Scheiner, Miiller and 
Kempf were found to be heartily in favor of the proposed journal and 
ready to promise their support and co5peration. In a plan of publica- 
tion formulated at Berlin it was decided that five Associate Editors be 
chosen to represent Germany, Great Britain, France, Italy and 
Sweden on the editorial staff, for it was felt from the first that unless 
the journal were made truly international in character it could not be 
a success. Professor Vogel readily consented to be the Associate 
Editor for Germany. In subsequent visits to Rome, Paris and London 
the plans of the journal were discussed at length with Professor P. 
Tacchini, Professor M. A. Cornu and Dr. William Huggins. Every- 
where the most cordial assurances of support and cooperation were 
received, and before my return to America the general plan of the 
journal had been decided upon, and the European members of the 
Board of Associate Editors chosen as follows : Professor M. A. Cornu, 
Ecole Polytechnique, Paris; Professor N. C. Dun6r, Astronomiska 
Observatorium, Upsala ; Dr. William Huggins, Tulse Hill Observa- 
tory, London ; Professor P. Tacchini, R. Osservatorio del CoUegio 
Romano, Rome ; Professor H. C. Vogel, Astrophysikalisches Observ- 
atorium, Potsdam. Subsequently five Associate Editors were secured 



82 MINOR CONTRIBUTIONS AND NOTES 

in the United States : Professor C. S. Hastings, Yale University: Pro- 
fessor A. A. Michelson, University of Chicago ; Professor E. C. Pick- 
ering, Harvard College Observatory ; Professor H^ A. Rowland, Johns 
Hopkins University ; Professor C, A. Young, Princeton University. 

Professor James E. Keeler, of Allegheny Observatory, whose asso- 
ciation in the editorial management of Astronomy and Astro-Physics 
had done so much for that journal, agreed to join the writer in editing 
The Astrophvsical Journal. Professor Henry Crew, of North- 
western University, and Professor Joseph S. Ames, of Johns Hopkins 
University, will continue the valuable work they have hitherto carried 
on in connection with Astronomy and Astro-Physics as Assistant 
Editors of the new journal, and Professor F. L. O. Wadsworth, 
of the University of Chicago, Professor Edwin B. Frost, of Dart- 
mouth College, and Professor W. W. Campbell, of the Lick 
Observatory, have promised to assist in the same capacity. In 
addition to this exceptional editorial cooperation — in itself quite suffi- 
cient to make The Astrophvsical Journal truly international in 
character — we are fortunate in having promises of assistance from 
many astronomers and physicists in Europe and America. 

It must not be supposed that The Astrophvsical Journal will deal 
only with the astronomical applications of the spectroscope. On the 
contrary, the scope of the Journal will be quite as broad as that of 
Astronomy and Astro-Physics has been, for while papers dealing only 
with questions of celestial mechanics and measures of the positions of 
the heavenly bodies will not fall within it, they will be replaced by 
articles treating of laboratory researches closely allied to the investiga- 
tions of astronomical physics. Drawings, photographs, descriptions and 
theories of the Sun, Moon, planets, satellites, comets, shooting stars, star 
clusters, nebulae and the Milky Way will all be considered as coming within 
the scope of the new journal. So too will observations of variable stars, 
photometric determinations of stellar n\agnitude and planetary albedo, 
measurements of solar radiation and atmospheric absorption, observa- 
tions of the phenomena of lunar and solar eclipses, and the numerous 
applications of the spectroscope in astronomy. The importance of 
supplying a common place of publication for papers on both the obser- 
vatory and laboratory applications of physical methods of research has 
already been pointed out. For this purpose much space will be 
devoted to articles on wave-length determinations of the lines in solar, 
metallic, and gaseous spectra, bolometric and radiometric work, spec- 



MINOR CONTRIBUTIONS AND NOTES 83 

tral photometry, experiments on radiation and absorption, photo- 
graphic researches in the infra-red and ultra-violet, studies of the rela- 
tions of the lines in different spectra, interference and diffraction phe- 
nomena, and theoretical work in certain branches of optics, heat, 
electricity and other departments of physics. In pursuance of the plan 
which seems to have met with favor in Astronomy and Astro- Physics^ 
the series of papers on the modern spectroscope will be continued, and 
accompanied by articles on telescopes, heliostats, bolometers, photo- 
meters and other instruments and apparatus used in such investigations 
as those mentioned above. Astrophysical and spectroscopic observa- 
tories and laboratories will also be fully illustrated and described. 

Special attention will be given to the reproduction of the latest 
photographs of astronomical and physical phenomena. By reason of 
their relations with the observatories and laboratories of Europe and 
America, the editors will always have the best photographs at their 
disposal. 

Articles written in any language will be accepted for publication, 
but unless a wish to the contrary is expressed by the author, they will 
be translated into English. 

In the department of Minor Contributions and Notes^ subjects other 
than those named in the above list of topics, but belonging to closely 
related fields of investigation, may find a place. 

It IS intended to publish in each number a bibliography of astro- 
physics, in which will be found the titles of recently published 
astrophysical and spectroscopic papers. In order that this list may be 
as complete as possible and that current work in astrophysics may receive 
appropriate notice in other departments of the Journal, authors 
are requested to send copies of all papers on these and closely allied 
subjects to both Editors. 

These and other details of the plan of publication of The Astrophysi- 
cal Journal were decided upon at a meeting of the American members 
of the Board of Editors held in New York on November 2, 1894. Profes- 
sors Young, Pickering, Rowland, Michelson, Hastings, Keeler and 
Hale were present. It was voted that a meeting of the Board of Edi- 
tors be held annually, for the discussion of matters relating to the 
Journal. 

Astronomy and Astro-Physics has been purchased from Professor 
Payne by the University of Chicago, and The Astrophysical Journal 
will be practically a continuation of it in a slightly different form. 



84 MINOR CONTRIBUTIONS AND NOTES 

The annual subscription price (for ten numbers) is $4.00 for the 
United States, Canada and Mexico. In other countries of the Postal 
Union the price is 18 shillings. Subscriptions should be sent to The 
University of Chicago^ University Press Division^ Chicago, Illinois, 

All European subscriptions should be sent to the sole foteign agents, 
Wm. Wesley b* Son, 28 Essex St,, Strand, London. 

AH papers for publication and correspondence relating to contribu- 
tions should be addressed to George E, Hale, Kenwood Observatory^ 
Chicago, Illinois, 

George E. Hale. 

NOTE ON THE ARC-SPECTRUM OF COPPER. 

Messrs. Crew and Tatnall publish in Astronomy and Astro-Physics, 
I3t 740, a method for obtaining the arc-spectra of metals free from 
carbon lines/ and as a specimen they give the spectrum of copper 
between X=40oo and X=36oo. In this region they find 41 new lines, 
which have not previously been published as lines of copper. In our 
copper spectrum Professor Runge and I have recorded only those 
lines on the photographs which we were convinced belonged to 
copper. A good many other lines, not appearing on all the plates, 
especially not on the weaker ones, were omitted as doubtful. Among 
these I find 23 of the lines measured by Crew and Tatnall, and thus it 
becomes more probable that these 23 lines really belong to copper. 
1 give here a list of these lines as measured by us : 



3976.12 


6 very hazy 


3800.55 


5 hazy 


3695.42 


5 hazy 


3964.40 


6 very hazy 


3800.06 


6 hazy 


3685.05 


6 


3947.09 


6 hazy 


3797.29 


6 hazy 


3644.20 


5 hazy 


3933.20 


6 hazy 


3780.14 


6 very hazy 


3632.65 


5 hazy 


3881.80 


6 hazy 


3764.90 


6 very hazy 


3629.91 


6 


3817.45 


6 very hazy 


3721.76 


6 very hazy 


3610.86 


5 hazy 


3813.62 


6 hazy 


3720.84 


6 


3609.39 


5 


3803.62 


6 very hazy 


3699.19 


5 hazy 







The concordance of our measurements with those of Crew and 
Tatnall is very satisfactory, considering that all the lines are very faint 
and hazy, and that only a few plates have been measured in both 
cases. Of the line 3619.52 Crew and Tatnall say : "surely not copper." 
This is, in fact, a very strong line of nickel. The line at 3961.64 is 

' It is certainly a great advantage to avoid the carbon lines, but I fear that 
Messrs. Crew and TatnalKs method will be of use only in cases where the high temper- 
ature of the carlx>n arc is not needed. 



MINOR CONTRIBUTIONS AND NOTES 85 

perhaps the strong line of aluminium. We have never observed the 
other 16 lines of Crew and Tatnall, but this is, of course, no reason for 
attributing them to impurities, as a stronger current always brings out 
new lines. 

H. Kayser. 

ON DETERMINING THE EXTENT OF A PLANET'S 
ATMOSPHERE. 

The early observations of the spectrum of Mars^ made between 
1867 and 1877, lead to substantially the same conclusion, viz.: the 
atmosphere of Mars is in a general way similar to our own. So far as 
I know, none of the observers formed, from the purely spectroscopic 
evidence, an estimate of the extent of the Martian atmosphere in terms 
of the extent of our own. The Moon and Mars were observed at equal 
altitudes, or in some cases with the Moon slightly lower than Mars. 
One observer saw one strong line and some faint lines in the planet's 
spectrum which were absent from the lunar spectrum. Another 
observer noted critical atmospheric and vapor bands in both spectra, 
but they were weaker in the lunar than in the Martian spectrum. A 
third observed bands in both spectra, but those seen in the lunar spec- 
trum were fewer and narrower than those seen in the planet's spectrum. 
Thus the three observers agreed that the critical bands were stronger 
in the spectrum of the planet than in that of the Moon, and for the 
reason that in the former case the light had passed twice either 
partially or completely through Mars' atmosphere. Now if it be true 
that the critical bands are stronger in one spectrum than in the other, 
and, further, if Mars' atmo^here is similar to our own, it ought to be 
possible to equalize the critical bands in the two spectra by observing 
the Moon's spectrum when that body is considerably lower in the sky 
than Mars is. If we could find those (unequal) altitudes of the two 
bodies such that the two spectra were equalized, we could at once 
compute the relative quantities of our atmosphere and aqueous vapor 
passed through by the light from the two bodies. The difference of 
those relative quantities would be the relative quantity of atmosphere 
and aqueous vapor passed through on Mars by that planet's light, and 
would represent considerably more than the maximum extent of Mars' 
atmosphere. This method would be valuable not only because it 
would enable us to form an estimate of the extent of the atmosphere, 
but it would also furnish a most excellent test of the delicacy of the 



86 MINOR CONTRIBUTIONS AND NOTES 

spectroscopic observations themselves. It seems to me this method is 
well worth applying. In considering the physical problems relating 
to Mars it is important to know whether or not there is a Martian 
atmosphere ; but it is no less important to know whether that atmosphere 
is very extensive or very thin. 

W. W. Campbell. 
Mt. Hamilton, November, 1894. 



THE CHICAGO ACADEMY OF SCIENCES. 

SECTION OF MATHEMATICS, ASTRONOMY AND PHYSICS. 

The regular monthly meeting was held December 11, at the Com- 
merce Club, Auditorium building ; Professor G. W. Hough, President, 
in the chair. After the transaction of routine business, the President 
introduced Dr. T. J. J. See, who presented a paper on ** Helmholtz's 
Theory of the Heat of the Sun." 

The speaker began by referring to Helmholtz's paper in the Philo- 
sophical Magaune iox 1856, which he said appeared to be very little 
known to astronomers. In this paper the illustrious physicist had 
given the formulae for computing the amount of heat developed by the 
condensation of the Sun, but had not shown how to derive the formulae. 
The speaker proceeded to determine the potential of a sphere upon 
itself, and to derive the formulae given by Helmholtz. It was shown, 
on reducing the formulae to numbers, that the condensation of the 
solar nebula from infinity had produced enough heat to raise the tem- 
perature of a mass of water equal to the Sun to 2 7, 000,000 ^C. As 
Pouillet had found by experiment that the heat annually lost by the 
Sun would raise the temperature of such a mass i.25^C, it follows that 
the age of the Sun cannot surpass 2 1,600,000 years, if its radiation has 
been uniform at the present rate. The speaker did not believe the 
radiation had been uniform, and hence was of the opinion that the 
Sun was much older than this result indicated. He proceeded to 
show that a contraction of -nriinr P^^ ^^ ^^ present radius would 
maintain the radiation for 2180 years ; this corresponds to a change of 
35 meters per year in the radius, and would of course be insensible for 
ages. It appeared that the condensation from infinity to the orbit of 
Neptune had produced only ^^^ part of the heat subsequently devel- 
oped. All these results were true on the supposition that the Sun is 
homogeneous, but in the actual case heterogeneity will considerably 



MINOR CONTRIBUTIONS AND NOTES 87 

modify the results ; the general effect of heterogeneity being to increase 
the total heat already lost by the Sun, and to lengthen its age by an 
unknown but considerable amount. In conclusion, the speaker regarded 
Helmholtz's theory as firmly established, but he took occasion to 
remark on the insufficiency of the older theories. He had found by 
calculation that if the Sun were pure carbon and pure oxygen in the 
right proportion to form carbon dioxide, the heat developed by the 
combustion of the entire mass would last only 1763 years. The 
meteoric theory was refuted on the ground that it required a sensible 
increase in the mass of the Sun, which would cause an acceleration in 
the mean motions of the planets — which has not been observed. In 
the discussion which followed. Professor Hough, Dr. Crew, Professor 
Chamberlin and Professor McNeill took part. 

Professor Chamberlin discussed the bearing of the paper on geolog- 
ical theories, and thought the time given for the age of the Sun was 
much too short, as geological phenomena seemed to require at least 
100,000,000 years for the age of the Earth. 

Dr. See pointed out that the heterogeneity of the Sun must very 
considerably increase its past duration, and especially if it radiated 
more slowly than at present. He thought the solar system was likely 
to be more than 100,000,000 years old, and regarded Lord Kelvin's 
estimate as a fair approximation. 

Dr. Crew thought modern estimates of the radiation would consid- 
erably modify the results of Pouillet. After considerable informal 
discussion, in which a number of members of the Academy partici- 
pated, the meeting adjourned. o*. 



WOLSINGHAM OBSERVATORY. 
CIRCULAR NO. 4 1. 

A very red 8 mag. IV type, not in DM., was found here last night 
at R. A. if 54".3; Dec. +58° 14' (1900)- 

T. E. EspiN. 
November 30, 1894. 



Reviews. 

On t/ie Spectrum of the Electric Discharge in Liquid Oxygen, Air 
and Nitrogen. Liveing and Dewar. Phil, Mag,, August, 
1894. 
In these experiments sparks were passed through small layers of 
the liquified gases, one or both electrodes being immersed. When 
both electrodes were immersed there was a continuous spectrum, 
probably due to the heated particles which left the surface of the 
electrodes, and also a line and band spectrum due to the liquid itself. 
The lines were faint and few unless a Leyden jar was put in the circuit. 
These lines could nearly all be identified with the known lines of the 
gases. When one electrode was out of the liquid, more bands appeared, 
and various interesting changes took place. The most noteworthy obser- 
vation made was that when sparks were passed through oxygen liquid 
and vapor, one electrode only being immersed, a band appeared between 
X 5530 and X 5610 ; and if a Leyden jar was put in the circuit, this 
band contracted to a line at about X 5572. The wave-length of the 
Aurora line is X 5571.6; and the conditions of temperature and 
pressure in these experiments must have been somewhat similar to 
those under which the Aurora appears. This points, of course, to the 
probability of the Aurora line being due to the oxygen of our atmos- 
phere. J. S. A. 

On Variations observed in the Spectra of Carbon Electrodes, and on 
the Influence of one Substance on the Spectrum of another. 
W. N. Hartley. Proc, R, 5., 55, No. 334. 

In this paper Professor Hartley reiterates his belief that the 
so-called " cyanogen " bands are in reality due to the element carbon, 
not to a compound of carbon and nitrogen. He thinks that the differ- 
ences between the "cyanogen" bands and the "carbon" bands are not 
too great to be accounted for perfectly by the variations in the external 
conditions under which the discharge passes between the carbon poles. 
He gives instances of many alterations produced by surrounding 
vapors and gases, and mentions the remarkable variations in intensity 

88 



REVIEWS 89 

of the lines of certain substances, produced by the addition of certain 
salts. These changes are undoubtedly real, and are most important ; 
but it seems difficult to explain by means of them the enormous differ- 
ences between the " cyanogen " and the *' carbon " bands. In fact the 
arguments as to their nature which were summed up by Kayser and 
Runge, and which seem to show that we have in reality two distinct 
substances, are in no wise weakened by this article. It might be well 
to remark here that there is at the present time no more fruitful field 
open to research than that of the study of the influence of the presence 
of one substance upon the spectrum of another. J. S. A. 



Flame Spectra at High Temperatures. II. and III. W. N. 
Hartley. Prac, R. 5., 56, No. 337. 

The first of these papers gives an abstract of a series of experiments 
upon the flame spectra of manganese and its compounds. The leading 
features of the spectra of the element and of its oxide are the same. 
One difiference, however, should be noted. In the group of lines about 
^ 403O1 the metal itself shows two bands closely adjacent, while the 
oxide gives what appears to be one band with a reversal down the 
middle, having the red side of the band sharp and strong, but the 
violet side very dififuse. This observation evidently has important 
bearing on the interpretation of Sun-spot spectra, where shadings occur 
on one side only of the metallic lines. 

In the second paper are given some important observations on the 
phenomena, spectroscopic and chemical, of the Bessemer process. By 
far the most interesting fact noted is the appearance of certain 
hydrogen lines in the flame emitted during the first period of the 
"blow." Hartley records his results as follows: "During the first 
period : The lines of the alkali metals, sodium, potassium, and lithium, 
are seen unreversed on a bright continuous spectrum caused by carbon 
monoxide. The Ha line of hydrogen, and apparently the Hfi line, 
were seen reversed during a snowstorm." Watts had previously 
observed the Ha line of hydrogen in the Bessemer flame during wet 
weather. If these observations can be accepted as definite, it would 
seem that the line spectrum of hydrogen can be produced, without the 
help of an electric discharge, at a temperature of about 1500^ C. An 
unpublished observation of Professor Rowland's is of interest in this 
connection. He has noted at least once that in the metallic arc-spectra 



90 REVIEWS 

a line appeared almost, if not quite, in the exact position of the By line. 

He has not felt convinced that it was not an " impurity " line, y q . 

J. o. A. 

Beitrage zur Kenntfdss der Litdenspectren. J, R. Rydberg. Wied. 
Ann., 50 (1893); 52 (1894). 
In these contributions to our knowledge of the line spectra of the 
elements, Rydberg calls attention to the similarity between various 
families of these, emphasizing particularly the importance of the 
numerical value of the differences between the wave-numbers of the 
consecutive lines in the same series. From the analogy between the 
spectra of certain elements, he predicts lines of definite wave-length in 
the spectrum of one or the other element. He has given special atten- 
tion to the lines of the spectra of calcium and strontium ; and has made 
a more complete study of the grouping of these lines into series than 
has ever been done before. He seems to hope ultimately to be able to 
identify all the lines with series obeying the same law as the hydrogen 
lines. He says that he thinks he has given additional reasons for believ- 
ing that there is only a single system of vibrations, and that all the lines 

of any one spectrum can be comprised in a single formula. t c a 

J. o. A. 

Beitrage zurKenntniss der Unienspectren. H. Kayser u. C. Runge. 
Wied, Ami,, 52 (1894). 

Under this title, the authors describe certain investigations prompted 
by the suggestion of Rydberg in his first series of Beitrage, They 
remark that the lines of the spectrum of magnesium which Rydberg 
thought formed a new series, cannot do so as they are so different 
physically. This emphasizes a warning to everyone who studies tables 
of wave-lengths instead of the spectra themselves. Kayser and Runge 
find, by careful study, certain additional lines and groups in the spectra 
of strontium, calcium, zinc and cadmium ; all of which Rydberg had 
predicted from analogy. In order to see plainly one of the strontium 
groups, which came under the cyanogen band at X 3800, Kayser and 
Runge weakened the effect of the carbon poles by passing a stream of 
CO, between them. When this was done, with strontium in the poles, 
a triplet was clearly photographed at X 3865.59, 3807.51, 3780.58. 

Although many lines predicted by Rydberg were found, some 
could not be, probably owing to lack of dispersion or to weakness of 
intensity. J. S. A. 



REVIEWS 91 

Ueber die Spectra von Zimi, BUi, Arsen^ Antifnati, Wismuth, H. 
Kayser u. C. Runge. Wied. Ann., 52 (1894). 
This paper gives the result of the study of the arc-spectra of the 
elements named, and is a continuation of the authors' previous work on 
the spectra of the elements. Many new lines were discovered, and many 
important differences between the arc and spark-spectra are noted. 
These points of difference are extremely valuable. The authors natur- 
ally looked diligently for any series or regular arrangements of the 
lines, such as occur in the spectra of the elements of the first three of 
Mendelejeff's groups. Although they found nothing like these, they 
did discover that in the spectra of each of the elements studied there 
were certain groups of lines such that there was a constant difference 
between the wave-numbers of a line in one group and a line in the 
other. The lines in the groups thus correspond ; and in one spectrum 
there may be as many as six groups, any two of which correspond. 
The lines forming any one group, however, do not seem to be con- 
nected by any mathematical or physical law. The number of lines 
which go to form these groups is too great to permit one to attribute 
them to chance, as Kayser and Runge prove quite conclusively. 

J. S. A. 

Preliminary Report on the Results Obtained with the Prismatic 
Camera during the Total Eclipse of the Sun, April 16, i8gj. 
J. Norman Locker. (Abstract.) 

TTu Total Solar Eclipse of April 16, i8gj. Report on Results Obtained 
tvith the Slit Spectroscopes. E. H. Hills. 

The initial number of Volume 56 of the Proceedings of the Royal 
Society contains two reports on the results of the English spectro- 
scopic observations of the total solar eclipse of April 16, 1893. 

The first is an abstract of a preliminary notice by Lockyer on the 
results of the observations with the prismatic camera. This instrument, 
of six inches aperture, and provided with a prism of 45^ angle, was 
used by Fowler at the African station. For the sake of a comparison 
of the results, another instrument — a spectroscope, with two three-inch 
prisms of 60^, used in connection with a sidprostat — was sent to the 
Brazilian station. 

In all thirty-two plates were secured during totality, and twenty-two 
just before and after totality. With instruments of this kind, without 



92 REVIEWS 

slits, the spectral lines become circles, or parts of circles. The H and 
K lines are the most conspicuous ones seen on the plates, and in them 
the forms of the prominences are clearly shown. The ultra-violet 
series of hydrogen lines is also prominent, and numerous other lines 
are visible, among them (on isochromatic plates) the *' corona line," 
^^ ^5317* ^^ ^^ plates show a bright continuous spectrum from the 
inner corona. 

The second report is by E. H. Hills, and refers to the results 
obtained with the slit spectroscopes, two of which were used at each 
station. Only one plate was taken with each instrument, the exposure 
being for three minutes and fifty seconds. The plates taken in Brazil 
were not successful, but a large number of lines were photographed at 
the African station. The plates were backed with a solution of asphalt 
in benzole in order to prevent the Feflection from the back surface of 
the glass. 

A list is given of the wave-lengths of seventy-one lines which are 
assigned to the prominence spectrum, and fifty-one lines credited to 
the corona, the "corona line " itself being put in the former category. 
The wave-lengths were determined by means of an interpolation curve 
based upon micrometric measures of the lines of the hydrogen series, 
with the " lines at X4215.3, X 4471.2 and the b group." The value of 
the results is greatly impaired by the fact that unreliable wave-lengths 
were assigned to the standards employed. The determinations by 
Huggins for the ultra-violet lines were the best at the time they were 
made, but they have been superseded by the measurements, with more 
powerful apparatus, by Cornu, Ames and others, and it seems very 
unfortunate that Angstrdm's scale should be employed in any new 
work. In consequence of this unhappy choice of standards the wave- 
lengths of the ultra-violet lines photographed at this eclipse differ in 
some cases as much as four tenth-meters from the values found by 
other observers of the prominence spectrum who base their results 
upon the Rowland scale. Much yet remains to be done in increasing 
the accuracy of our knowledge of the wave-lengths of the bright lines in 
the ultra-violet, but in the opinion of the reviewer the results under 
consideration will require re-reduction before they can be of service. 

E. B. F. 



Recent Publications. 



A LIST of the titles of recent publications on astrophysical and 
allied subjects will be printed in each number of The Astrophysical 
Journal. In order that these bibliographies may be as complete as 
possible, authors are requested to send copies of their papers to both 
Editors. 

For convenience of reference, the titles are classified in thirteen 
sections. 

1. The Sun. 

Brester, a., Jr. On Brester's Views as to the Tranquillity of the Solar 

Atmosphere. A. and A. 13, 849-856, 1894. 
Deslandres, H. On the Electric Origin of the Solar Chromosphere. 

Knowl. x7, 277, 1894. 
Deslandres, H. Recherches sur les mouvements de 1* atmosphere 

solaire. C. R. 1x9, 457> 1894. 
Flammarion, C. Sur la rotation des taches solaires. C. R. xxg, 

532, 1894. TAstr. X3, 42i-423» 1894. 
Hale, George £. On Some Attempts to Photograph the Solar 

Corona without an Eclipse. A. and A. 13, 662-688, 1894. 
Hazen, H. a. Sun-spots and Auroras. Am. Met. Jour, xx, 221-229, 1894. 
Hills, E. H'. The Total Solar Eclipse of April i6, 1893. Report on 

Results Obtained with the Slit Spectroscopes. Proc. R. S. 56, 20-26, 

1894. 
LocKYER, J. N0R.MAN. Preliminary Report on the Results Obtained 

with the Prismatic Cameras during the Total Eclipse of the Sun, 

April 16, 1893. Fhil. Trans. X85 A, 711-717, 1894. 
Wilson, W. £. and P. L. Gray. Experimental Investigations on the 

Effective Temperature of the Sun, made at Daramona, Strccte, County 

Westmeath, Ireland. Phil. Trans. X85 A, 361-396, 1894. 

2. The Solar Spectrum, Infra-Red, Visible, and Ultra-Violet. 
Langley, S. p. On Recent Researches in the Infra-Red Spectrum. 

Report Oxford Meet. B. A. A. S. 1894. Nat. 51, 12-16, 1894. 

3. Stars and Stellar Photometry. 

Barnard, E. E. and A. C. Ranyard. Structure of the Milky Way. 
Knowl. x7, 253. 1894. 

93 



94 RECENT PUBLIC A TIONS 

Chandler, S. C. Note on the Variable Star Z Hercults. A. N. 

»36, 331-332, 1894. 
Chandler, S. C. On a New Variable of the Algol-Type, 6442, Z 

Herculis. Astr. Jour. No. 328, 14, 125, October 20, 1894. 
Chandler, S. C. Ephemerides of Long-Period Variables for 1895. 

Astr. Jour. No. 327, 14, 118-119, October i, 1894. 
Editor of the Astr. Jour. Announcement as to Two Telescopic 

Variables. Astr. Jour. No. 330, 14, 144, November 23, 1894. 
Editor of the Astr. Jour. New Variable of the Algol-Type. 

Astr. Jour. No. 327, 14, 120, October i, 1894. 
Gore, J. E. Globular Star Clusters. Knowl. 17, 232, 1894. 
Gore, J. £. The Distance and Mass of the Binary Stars. Knowl. 

17, 271, 1894. 
Hartwig, Ernst. Ueber den neuen veranderlichen Stern Z Herculis. 

A. N. 136, 329-332» 1894. 
Pannekoek, a. Beobachtungen des neiien Veranderlichen Z Her- 
culis. A. N. 136, 221, 1894. 
Parkhurst, H. M. Stellar Photometry. A. and A. 13, 652-659, 1894. 
Parkhurst, J. A. Maxima and Minima of Long-Period Variables. 

Astr. Jour. No. 331, 14, 151, December 10, 1894. 
Plassman, J. Beobachtungen von Z Herculis in Warendorf. A. N. 

136. 333-334. 1894. 
Ranyard, a. C. Photographs of the Milky Way and Nebulae. 

Knowl. 17, 226, 1894. 
Reed, William Maxwell, Observations of Variable Stars. Astr. 

Jour. No. 330, 14, 137-141, November 23, 1894. 
Roberts, A. W. New Short-Period Variable. Astr. Jour. No. 327, 

14, 120, October i, 1894. 
Roberts, A. W. Variation of (3416) — Velorum and (5949) — Arae. 

Astr. Jour. No. 327, 14, 113-117, October i, 1894. 
Sawyer, Edwin F. On the Variability of 6442 Z Herculis. Astr. 

Jour. No. 329, 14, 135, November 9, 1894. 
Sawyer. Edwin F. On the Variable Star 6573 Y Sagiltarii. Astr. Jour. 

No. 328, 14, 127-128, October 20, 1894. 
Yexdell, Paul S. On the Variability of DM. + I5" 33ii. Astr. 

Jour. No. 328, 14, 125, October 20, 1894. 
Yexdell, Paul S. Observed Maxima of Long-Period Variables, with 

Note on a Suspected New Variable. Astr. Jour. No, 328, 14, 121-122, 

October 20, 1894. 
Yexdell, Paul S. On a New Variable of Short Period. Astr. Jour. 

No. 331, 14, 150, December 10, 1894. 
Yexdell, Paul S. Observations of Suspected Variables. Astr. Jour. 

No. 329, 14, 133-134, November 9, 1894. 



RECENT PUBLIC A TIONS 9 5 

Stellar Spectra, Displacements of Lines and Motions in the 

Line of Sight. 
Belopolsky, a. Das Spectrum von 8 Cephei. A. N. 136, 281- 

284. 1894. 
ESPIN, T. E. The Spectrum of a Herculis. A. and A. 13, 651, 1894. 
LoCKYER, J. Norman. The Spectrum Changes in P Lyrae. Proc. 
R. S. 56. 278^285, 1894. 

Planets, Satellites and their Spectra. 

Barnard, E. E. The Great Red Spot and other Markings on Jupiter. 

A. and A. 13, 736, 1894. 
Barnard, E. E. The Form of the Disk of the III Satellite and Phe- 
nomena of the Occultation of a Satellite of Jupiter. A. and A. 13, 

821, 1894. 
Bigourdan, G. Disparition de la tache polaire australe de Mars. 

C. R. XX9, 633,840, 1894. 
Cerulli, V. Una macchia sul lembo nord di Marte. A. N. 136, 

223, 1894. 
Campbell, W. W. The Spectrum of Mars. Pub. A. S. P. 6, 228, 

1894. A. and A. 13, 752-760, 1894. 
Douglass, A. E. The Polar Cap of Mars. A. and A. X3, 738, 1894. 
Elger, T. Gwyn. The Central Equatorial Region of the Moon. 

Khowl. 17, 276, 1894. 
Elger, T. Gwyn. Selenographical Notes. Obs*y 17, 266, August, 1894 ; 

332, October, 1894; 357, November, 1894. 
Flamharion, C. Les neiges polaires de Mars. C. R. xzg, 786-791, 

1894. 
Flam MARION, C. Sur ies pdles de rotation de V£nus. C. R. 119, 

670, 1894. 
Flammarion, C. Observations de la plan^te Mars, faites & 1' Obser- 

vatoire de Juvisy. V Astr. x3, 376-385 ; 408-413, 1894. 
Franz, J. Der Einfluss der Phase auf die scheinbare Lage von Mdsting 

A. A. N. 136, 321-330, 1894. 
Gaudibert, C. M. Photographie lunaire. Clavius. TAstr. 13, 413- 

419. 1894. 
Gruey, L. J. Observation de 1' Eclipse partielle de Lune 1894, Septem- 

bre 14, faite k I'Observatoire de Besan^on. A. N. 136, 285-286, 1894. 
Lowell, Percival. Mars. A. and A. 13, 538-553, 645-650, 740- 

741, 814-821, 1894. 
Martu, a. Ephemeris for Physical Observations of Jupiter, 1894-5^ 

M. N. 54. 562, 594, 1894. 
Maunder, E. W. The Canals of Mars. Knowl. xy, 249, 1894. 



96 RECENT PUBLIC A TIONS 

Pickering, W. H. Schiaparelli*s Latest Views Regarding Mars. A. 

and A. X3, 632-635, 1894. 
RiSTENPART, F. and MiE, G. Beobachtung der partiellen Mondfin- 

sterness, 1 894, September 1 4, zu Karlsruhe. A. N. 136, 283-286, 1 894. 
ScHiCBERLE, J. M. The Region of Lacus Solis on Mars. A. and A. 

X3, 644. 1894. 
ScHiAPARELLi, G. The Planet Mars. A. and A. 13, 635-640, 714- 

722, 1894. 
Trouvelot, E. L. Passage de Mercure devant le Soleil. C. R. izg, 

842, 1894. 
Williams, A. Stanley. Notes on Mars, 1894. Obs'y 17, 319-321, 

October, 1894; 347-349f November, 1894. 
Young, C. A. Transit of Mercury, 1894, November 10. Astr. Jour. 

No. 330, 14, 143, November 23, 1894. 

6. Comets, Meteors and their Spectra. 

Barnard, £. £. Photographs of a Remarkable Comet. A. and A. 

X3. 7891 1894. 
Denza, p. F. Les ^toiles filantes observ^es en Italie au mois d*aoOt, 

1894. C. R. 119, 508, 1894. TAstr. Z3, 419, 1894. 
Maltezos, C. Sur la chute des bolides et a^rolithes tomb^s demi^re- 

mem en Gr^ce. C. R. xig, 500, 1894. TAstr. 13, 386, 1894. 
Wesley, W. H. The Comet on the Eclipse Photographs of 1893. 

Obs'y 17, 349-353. November, 1894. 

7. Nebulae and their Spectra. 

Barnard, £. E. Photograph of M. 8 and the Trifid Nebula. A. and 
A. 13, 791 » 1894. 

Barnard, £. £. Photograph of Swift's Nebula in Monoceros. N. G. C. 
2237. A. and A. 13, 642-6441 1894. 

Barnard, E. E. The Great Photographic Nebula of Orion, Encirc- 
ling the Belt and Theta Nebula. A. and A. 13, 81 1-814, 1894. 

Barnard, E. E. On the Exterior Nebulosities of the Pleiades. A. 
N. Z36, 196, 1894. A. and A. 13, 768-770, 1894. 

Easton, C. The Great Nebula in Andromeda. Nat. 50, 547, Octo- 
ber 4, 1894. 

LoCKYER, J. Norman. On the Photographic Spectrum of the Great 
Nebula in Orion. Proc. R. S. 56, 285, 1894. 

8. Terrestrial Physics. 

BiGELOW, F. H. Inversion of Temperatures in the 26.68-day Solar 
Magnetic Period. Am. Jour. [3] 48, 435- 45 't 1894. 



RECENT PUBLIC A TIONS 97 

CULVERWELL, EDWARD, P. A Criticism of the Astronomical Theory 
of the Ice Age. Nat. 51, 33-34» November 8, 1894. 

CuLVERWELL, EDWARD P. A Mode of Calculating a Limit to the Direct 
Effect of Great Eccentricity of the Earth's Orbit on Terrestrial 
Temperatures, Showing the Inadequacy of the Astronomical Theory 
of Ice Ages and Genial Ages. Phil. Mag. [5] 38, 541-552. 
December, 1894. 

9. Experimental and Theoretical Physics. 

Abney, W. de W. Measurement of Color Produced by Contrast. 

Proc. R. S. 56, 221-329, 1894. 
deMuynck, R. Ueber die Brechungsexponenten von wSsserigen 

Cadmiumsalzl5sungen. Wied. Ann. No. 11, 53, 559-563, 1894. 
Nichols, E. L. and Mary L. Crehore. Studies of the Lime Light. 

Phys. Rev. a, 1 61-170, 1894. 
Paschen F. Ueber die Dispersion des Steinsalzes im Ultraroth, 

Wied. Ann. No. 10. 53, 337-342, 1894. 
Paschen, F. Ueber die Dispersion des Fluorits im Ultraroth. Wied. 

Ann. No. 10. 53, 30>-333» 1894. 
PiCTET, Raoul. l^tude sur le rayonnement aux basses temperatures ; 

applications & la th^rapeutique. Arch, de Geneve 3a, 233-254, 465- 

480, 1894. 
Rubens, H. Prilfung der Ketteler-Helmholtz'schen Dispersionsformel. 

Wied. Ann. No. 10, 53, 267-286, 1894. 
Sharp, C. H. and W. R. Turnbull. A Bolometric Study of Light 

Standards. Phys. Rev. a, 1-35, 1894. 
Salomons, David. On some Phenomena in Vacuum-Tubes. Proc. 

R. S. 56, 229-250. 1894. 
Thomas, L. Sur la constitution de Tare ^lectrique. C. R. xxg, 728, 1894. 

10. The Spectra of the Elements. 

Crew, Henry and R. Tatnall. On a New Method for Mapping the 
Spectra of Metals. Phil. Mag. 38, 379-386, 1894. A. and A. 13, 

741-747. 1894. 
Ewan, Thomas. On the Absorption Spectra of Dilute Solutions. 

Proc. R. S. 56, 286, 1894. 
Hartley, W. N. Flame Spectra at High Temperatures, Part H. 

The Spectrum of Metallic Manganese»of Alloys of Manganese, and of 

Compounds Containing that Element. Proc. R. S. 56, 192, 1894. 
Hartley, W. N. Flame Spectra at High Temperatures. Part IIL 

The Spectroscopic Phenomena and Thermo-Chemistry of the Besse> 

mer Process. Proc. R. S. 56, 193-196, 1894. 



98 RECENT PUBLIC A TIONS 

Hasselberg, B. On the Line Spectrum of Oxygen. Wied. Ann. 

No. 8, 1894. A. and A. 13, 760-763, 1894. 
Lewis, E. P. and E. S. Ferry. The Infra-Red Spectra of Metals. 

J. H. Univ. Circ. No. 12. A. and A. 13, 747-752, 1894. 
Paschen, F. Bolometrische Arbeiten. Wied. Ann. No. 10, 53, 287-r 

300, 1894. 
Paschen, F. Die genauen Wellenlangen der Banden des ultrarothen 

Kohlensaure und Wasserspectrums. Wied. Ann. No. 10, 53, 334- 

336, 1894. 

ti. Photography. 

Meldola, R. The Optics of Photography. Review of Vogers Handbuch 
der Photographic, IL Theil. Nat. 50, 589-590, October 18, 1894. 

12. Instruments and Apparatus. 

Deslandres, H. Sur Tenregistrement de la chromosphere et de la 
photosphere du Soleil par la m^thode des sections successives. Bull. 
Astr. IX, 425-426, October, 1894. 

Elkin, W. L. Instrument for the Photography of Meteors. A. and 
A. 13, 626, 1894. 

Howe, H. A. The 20-inch Equatorial of the Chamberlin Observa- 
tory. A. and A. 13, 709-714, 826-830, 1894. 

Nichols, E. L. A New Form of Spectrophotometer. Phys. Rev. a, 
138, 1894. 

PULFRICH, C. Ueber eine neue Spektroskop-Konstruktion. Z. f. 
Instrum. 14, 354-364, 1894. 

Riefler, S. Beschreibung des Echappements mit vollkommen freien 
Pendel. Z. f. Instrum. 14, 346, 1894. 

SiRKS, J. L. On the Astigmatism of Rowland's Concave Gratings. 
Ver. K. Akad. v. Wetens, a. A. and A. 13, 763-768, 1894. 

Wadsworth, F. L. O, Improved Form of Interrupter for Large 
Induction Coils. Am. Jour. [3] 48, 496-501, 1894. 

Wadsworth, F. L. O. Eine neuer Spektroskopspalt mit Doppelbeweg- 
ung. Z. f. Instrum. 14, 364, 1894. 

Wadsworth, F. L. O. Some New Forms of Double Motion Mechan- 
ism. A. and A. 13, 527-538, 1894. 

Wadsworth, F. L. O. A Simple Method of Mounting an Equatorial 
Axis on Ball Bearings. A. and A. 13, 723-728, 1894. 

Wadsworth, F. L. O. Fixed-Arm Spectroscopes (The Modem Spec- 
troscope, IX.). Phil. Mag. 38, 337-351. October, 1894; A. and A. 

13, 835-849. 
. The Buckingham Works, York (Cooke's Equatorial 

Mountings). Engineering 58, 660-665, 1894. 



RECENT PUBLIC A TIONS 99 

13. General Articles, Memoirs and Serial Publications. 

Blanshard, C. T. New Element in the Sulphur Group. Nat. 50, 

571, October 11 , 1 894. 
Dreyer, J. L. £. Astronomical Spectroscopy ; A Review o! Frost's 

translation of Die Spectralanalyse der Gestime, Nat. 50, 565-567, 

October 11, 1894. 
Dyson, F. W. Note on Professor Turner's paper on the Reduction of 

Measures of Photographic Plates. M. N. 54, 573, 1894. 
Hale, George £. The Astrophysical Journal. A. and A. 13, 831-835, 

1894. 
Keeler, James £. Professor Frost's Translation and Revision of Die 

spectralanalyse der CesHme. A. and A. 13, 668, 1894. 
Lohse, O. Planetographie. Eine Beschreibung der im Bereiche der 

Sonne zu beobacht. K5rper, 8vo. x, ix+192, 15 cuts. Weber, 

Leipzig, 1894. 
Mendenhall, T. C. Measurements of Precision ; An address delivered 

at the Johns Hopkins University, Baltimore. Nat. 50, 584-587, 

October 11, 1894. 
Michelson, a. a. Determination experimentale de la valeur du 

m^tre en longueurs d'ondes lumineuses. Mem. du Bur. Int. des 

Poids et Mesures xx, 1894. 
MULLER, G. and P. Kempf. Publicationen des Astrophysikalischen 

Observatoriums zu Potsdam, Nr. 31. Neunter Band. Photometrische 

Durchmusterung des ndrdlichen Himmels, enthaltend alle Sterne der 

B. D. bis zur Grdsse 7.5. Theil I., Zone 0° lis + 20" Declination. 

Potsdam, 1894. 
Strehl, Karl. Die Theorie des Fernrohrs auf Grund der Beugung 

des Lichts. I. Teil, viii.4-136, i plate, 8vo, 1894. J. A. Barth, 

Leipzig, 



NOTICE. 

The scope of The Astrophysical Journal includes all investigations of 
radiant energy, whether conducted in the observatory or in the laboratory. 
The subjects to which special attention will be given are photographic and 
visual observations of the heavenly bodies (other than those pertaining to 
"astronomy of position*'); spectroscopic, photometric, bolometric and radio- 
metric work of all kinds ; descriptions of instruments and apparatus used in 
such investigations ; and theoretical papers bearing on any of these subjects. 

In the department of Minor Contributions and Notes subjects may be 
discussed which belong to other closely related fields of investigation. 

It is intended to publish in each number a bibliography of astrophysics, 
in which will be found the titles of recently published astrophysical and 
spectroscopic papers. In order that this list may be as complete as possible, 
and that current work in astrophysics may receive appropriate notice in other 
departments of the Journal, authors are requested to send copies of all 
papers on these and closely allied subjects to both Editors. 

Articles written in any language will be accepted for publication, but 
unless a wish to the contrary is expressed by the author, they will be translated 
into English. If a request is sent with the manuscript twenty-five reprint 
copies of each paper, bound in covers, will be furnished free of charge to the 
author. Additional copies may be obtained at cost price. No reprints can 
be sent unless a request for them is received before the Journal goes to 
press. 

The Editors do not hold themselves responsible for opinions expressed by 
contributors. 

The Astrophysical Journal will be published on the first day of every 
month except July and September. The annual subscription price for the 
United States, Canada and Mexico is $4*00 ; for other countries in the Postal 
Union it is 18 shillings. Correspondence relating to subscriptions and adver- 
tisements should be addressed to The University of Chicago^ University Press 
Division^ Chicago^ IIL 

Wm, Wesley &* Son, 28 Essex St., Strand, London, are sole foreign agents, 
and to them all European subscriptions should be addressed. 

All papers for' publication and correspondence relating to contributions 
and exchanges should be addressed to George E, Hale, Kenwood Observatory, 
Chicago, III. 



100 



PLATE VI 




ARTHUR COWPER RANYARD 



THE 

ASTROPHYSICAL JOURNAL 

AN INTERNATIONAL REVIEW OF SPECTROSCOPY 
AND ASTRONOMICAL PHYSICS 



VOLUME I FEBRUARY 189J number a 



ON A LENS FOR ADAPTING A VISUALLY CORRECTED 
REFRACTING TELESCOPE TO PHOTOGRAPHIC 
OBSERVATIONS WITH THE SPECTROSCOPE." 

By James E. Keeler. 

Although photography is continually becoming more impor- 
tant in its astronomical applications, the old method of eye obser- 
vation is still regarded as entitled to precedence when large tele- 
scopes are constructed, as we may see in the fact that all the 
very large refractors of modern times have visual object-glasses. 
Such instruments are sometimes adapted to photographic work 
by placing a third lens or photographic corrector over the 
visual objective, the effect of which is to change the chromatic 
aberration in the desired manner, and at the same time to 
greatly shorten the focus. Other, and no doubt better methods 
have so far been applied only to comparatively small instru- 
ments. 

Thus the spectroscopist who has a very large telescope at his 
disposal is compelled to work with a visual objective, and in 
attempting to photograph the spectra of stars, or other objects 

'Communicated, in a considerably modified form, to the Astronomical and 
Physical Society of Toronto. 

lOI 



102 JAMES E. KEELER 

having a small angular magnitude, he encounters many difficul- 
ties. When a slit spectroscope is used in connection with a 
refracting telescope, it is evident that the spectrum of a star 
will be linear only at that point for which the focus of the 
refractor is a minimum. The spectrum will not even then be 
strictly linear unless the lenses of the spectroscope are also achro- 
matized for the same rays, although any departure from this lat- 
ter condition will produce a relatively small effect. With a vis- 
ual objective the spectrum is, therefore, linear at the brightest 
part; but the greatest photographic action is found near the 
hydrogen line //y, which is the line most used for photographic 
determinations of motion in the line of sight, and here the focus 
of the objective varies rapidly with the wave-length. In the 
case of the Lick telescope, for example, if the slit is in the focal 
plane for the Hy rays, it is 36"".8 outside the focus for Hfi 
and 3 3"". 3 inside the focus for Hh, The cone of rays corre- 
sponding to the former line has a diameter of i""^.9, where it 
is intersected by the slit-plate, and that corresponding to the 
Hi line has a diameter of i"".7. As the slit is only about one- 
fiftieth of a millimeter wide, very few of these rays can enter. 
The spectrum, instead of being linear, and of nearly uniform 
intensity between Hfi and Hh^ is very narrow at //y, widens 
rapidly on both sides of that line, and falls off very rapidly in 
intensity. Hence only an extremely short range of the spectrum 
can receive the proper exposure ; and not only is the extent of the 
photograph greatly reduced in this way b}' the chromatic aber- 
ration, but reliable estimates of the relative intensities of lines 
are made almost impossible. These difficulties are perhaps only 
fully realized by one who has actually attempted to make the 
photographs. 

Another difficulty of considerable practical importance arises 
from the fact that the eye must be used in guiding the telescope 
during an exposure, and that the Hy star-image, which must be 
kept within the slit, is but little brighter than the non-actinic 
expanded star-disks which surround it. This difficulty can he 
partly remedied by the use of blue glass, but even then it is 



A PHOTOGRAPHIC CORRECTING LENS IO3 

impossible to be certain that the image is properly centered, or 
even within the slit. 

It is, therefore, highly desirable to find some way of chang- 
ing the chromatic aberration of a visual telescope so as to unite 
the photographically active rays. We have seen that this end 
can be attained by the use of a specially corrected third lens, 
but the photographic corrector, besides being expensive, short- 
ens the focus so greatly that the spectroscope cannot be placed 
in the proper position. In the case of the Lick telescope the 
focus is inside the tube, about nine feet above the eye-end, where 
it is inaccessible for spectroscopic purposes. 

What is required, therefore, is a small lens, which, placed in 
the cone of rays from the large objective not very far above the 
eye-end, will effect the desired change in the chromatic aberration 
without greatly shortening the focus. Such a lens, it may be 
observed, would be quite useless for ordinary celestial photog- 
raphy, as the distortion at a short distance from the axis would 
be very great. For stellar spectroscopy, however, a field less than 
one millimeter in diameter is quite sufficient, and the question of 
distortion need not be considered. 

The problem, when it includes the object-glass, is a com- 
plicated one ; but starting with the color curve, which can easily be 
determined experimentally by means of the spectroscope, the 
approximate solution is very simple. Some form of correcting 
lens is, I believe, actually in use, but I have not been able to 
find anything in print on the subject, probably because it is only 
quite recently that the necessity for a device of this kind has 
arisen, and the following elementary consideration of the ques- 
tion may be of some interest.' It would hardly be worth while, 

' After the type of the present article had been set up, my attention was called to 
a brief account of a single correcting lens by Mr. Newall ('* Notes on some Photo- 
graphs taken with a Visual Telescope," At. N,^ 54, 373). The lens is placed five feet 
above the focus of the Cambridge 25*inch equatorial, and reduces the separation of the 
/^/9 and /^e foci from 1.5 inches to about two-tenths of an inch. The focal length of 
the combination is shorter by about 18 inches than that of the object-glass alone. This 
shortening of the focus, although regarded by Mr. Newall as an advantage, seems to 
me to be practically very objectionable in the general case of a large refracting tele- 
scope, for reasons which I have given farther on. • 



104 



JAMES E. KEELER 



for purposes of illustration, to make an accurate computation of 
the correcting lens, and I have employed the approximate for- 
mulae for lenses, in which the thickness and the spherical aberra- 
tion are neglected. I have also chosen for illustration the case 
of the Lick telescope, because its color curve is known (Plate VII, 
Fig. I ) .' The other data required are the refractive indices of 
crown and flint glass, and these I have taken (to four decimal 
places) from the tables of Hopkinson. The "dense flint" of the 
tables is nearly the same as the glass of a prism in my possession, 
which is slightly colored ; a lighter flint would probably be better, 
as the curvatures required are very moderate. The following table 
contains all the necessary data. In the column headed "color 
curve " are given the distances in meters of the different foci of 
the great telescope from a point on the axis one meter above the 
focus for the H^ line. 







COLOK CUBVB 


LiNB OF SpBCTRUM 


Hard Crown 


Dente Flint 


Dittanoe of Focus 
fionLens 


B 
Na 

E 

m 


I. 5136 
1.5146 
I.517I 
X.5203 
I. 5210 
I. 5231 
1.5280 
I .5309 


I. 6157 
I. 6175 
1.6224 
1.6289 
1.6302 
1.6347 
1.6453 
I. 6518 


I. 0000 

0.9939 
0.9886 

0.9905 
0.9914 
I. 0000 

1.0368 

I. 0701 



The lens may be placed in the cone of rays at any convenient 
point. We shall suppose that it is placed one meter above the 
focus for* the Hfi line, in which case its aperture must be (at 
least) S3"". 

In order to make the spectrum of a star-image linear at Hy 
it will be suflicient (at least for our present purpose) to unite 
the rays Hfi and Hi,* 

' Platted from the table on p. 162, Vol. II., Publications of ike Asironomicai Society 
of the Pacific, See also Publications of the Lick Observatory^ 3, 174. 
'Actually, the lower ray should be a litUe above H^, 



i - 




9000 000009 






o 
U 
-o 

B 
S 

s. 

E 

c 
U 

■? 

& 

3 

u 

o 
"o 
U 



H 
a. 



;£ 



«• •»» in 



"T" 

00 



c 



I 

u 

v. 

C 



s 
U 

o 
"o 
u 




t. 



15* 



d > 



t 

a 
U 



c 
U 



A PHOTOGRAPHIC CORRECTING LENS 105 

It will be interesting to first consider the case of a single 
lens. Since the Hi must be more deviated than the H/3 rays in 
order that both may meet at the same point on the axis, the lens 
must be convex ; further, since it is desirable to produce this 
dispersive effect with the least amount of refraction, the lens 
should preferably be of flint glass. 

Let u be the distance from the lens at which the rays are 
united, v and v' the distances from the lens of the original foci 
for Hp and Hi respectively, as given in the table. If the point 
whose distance is u is regarded as a source of light, the Hfi and 
Hi rays, after refraction by the lens, would appear to diverge 
from points whose distances are respectively v and v' , Hence 
u and V are conjugate foci for the Hfi rays, and u and v' are 
conjugate foci for the Hi rays. 

The general equation connecting the conjugate foci of a thin 
lens, its refractive index, and the radii of its surfaces, is 

\-z = ^- '<-'-> 

in which all lines measured from the lens toward the source of 
light (at u, in this case toward the focus of the objective) are 
positive.' As the shape of the lens, or relation between the 
radii of its surfaces, is not taken into account in these approx- 
imate formulae, we may for simplicity consider the surface turned 
toward the eye-end to be flat, and this form is easily seen on 
general principles to be very nearly the best with regard to 
spherical aberration. Placing r= oo , the formula becomes 

X I /A I 

V u s 

If the symbols in this formula are regarded as referring to 
the Hp line, we may write for the Hi line 
I I /i' — I 

and since u is the same in both cases, 

— H = — ,- -\ , or j- 

V s V s V V 

' According to this convention, the focal length of a convex lens is negative. 



io6 



JAMES E. KEELER 



From the table of data we have: /»= 1.6347, /i' = i.65i8, «'= 
1. 0000, t/' = 1.0701, and substituting these values we find for the 
radius of the lens, j= +0.261 1 meters. 

For the focal length of the lens (for the Hfi line) we find 
from the formula r 

/== — 0.4 1 1 3 meters. The focal length for the HI line is —0.4005 
meters. The distance u of the point at which the rays are 
united is given by i i i 

and is found to be 0.2915 meters. 

To obtain the form of the color curve, as modified by the 
correcting lens, we substitute successively for v the values in 
the last column of the table of data, and compute the correspond- 
ing values of u, using for each ray its appropriate index. The 
values of u so obtained are the ordinates of points on the modi- 
fied color curve. The results are given in the following table : 



Link 


M 


V 


u 






m 


m 


B 


I. 6157 


I. 0000 


0.2978 


Ha 


1-6175 


0.9939 


0.2966 


D. 


1.6224 


0.9886 


0.2945 


E 


1.6289 


0.9905 


0.2925 


^ 


1.6302 


0.9914 


0.2922 


m 


1.6347 


I. 0000 


0.2915 


Hy 


1-6453 


1.0368 


0.2910 


m 


I. 6518 


I. 0701 


0.2915 



The resulting color curve is shown in Fig. 2, which is drawn 
on the same scale as Fig. i. It shows a very great advantage 
over the latter, not only in the part of the spectrum considered 
in its determination, but in the lower spectrum as well. Never- 
theless the single lens is impracticable, for the following rea- 
sons : in the first place, the focus of the telescope is shortened 
about 0.7 meter, which with the ordinary construction of the 
eye-end of a telescope would make the focal plane inaccessible ; 



A PHOTOGRAPHIC CORRECTING LENS IO7 

in the second place, the convergence of the cone of rays is 
increased from 1:19 to ^-S-SS* which would necessarily also be 
the angular aperture of the collimator. With a given effective 
aperture the collimator would therefore have to be very short, a 
construction which is very disadvantageous in every respect. 
It would, moreover, be difficult or impossible to make a suffi- 
ciently good collimator lens with such a very large angular aper- 
ture. 

A single correcting lens placed two meters above the focal 
plane of the great objective would have an aperture of I05"*".3 ; 
its focal length would be i"^.59i, and it would shorten the focal 
length of the telescope by i"".ii4. The ratio of aperture to 
length of the transmitted cone of rays would be 1:8.42. A lens 
of the same kind placed close to the telescope objective becomes 
the ordinary photographic corrector. The greatest shortening 
of the focus is then produced, and at the same time the angle of 
the convergent cone of rays is the least possible. 

The condition that the focus of the telescope must not be 
greatly altered by the correcting lens is practically a very 
important one. With the double lens which will next be con- 
sidered it would optically be an advantage to shorten the 
focus until the convergence of the transmitted cone was 
about 1 : 15, as a considerable improvement in the modified color 
curve would result. This is, however, forbidden by practical 
difficulties of construction. To make the whole spectroscope 
movable through any considerable distance along the axis of 
the telescope, in order to follow the change of focus, would not 
be an easy matter in the case of a large instrument ; it could 
hardly be done without some sacrifice of stability, it would 
increase the cost of construction, and in any case it would be 
extremely inconvenient. On the other hand, to increase the 
range of the collimator-slide (already several inches) would 
sensibly diminish the stability of the collimator, on which the 
value of the spectroscope as an instrument of precision prima- 
rily depends. It is therefore a matter of practical importance to 
keep the focus of the telescope very nearly in the same place. 



I08 JAMES E. KEELER 

The color curve can be transformed in the required manner 
without changing the position of the focus, by means of a 
double lens. If we increase the curvature of the convex flint 
lens hitherto considered, and neutralize the excess of dispersive 
effect by a concave lens of crown glass, the superior refractive 
power of the latter will give the combination a greater focal 
length than that of the equally dispersive single lens, and it 
will therefore not converge the rays from the objective so 
strongly. By suitably combining the lenses we can leave the 
position of the focus for any given ray unchanged. 

In determining the focal lengths of the component lenses we 
shall as before assume that the lens is placed one meter above 
the focal plane for the Hfi line, and use only approximate for- 
mulae. 

For thin lenses in contact, the relation between the conju- 
gate foci and the focal lengths of the lenses is expressed by the 
equation i__l— i ri 

In this, and in the following formulae, quantities relating to 
the crown lens will be indicated by the subscript i , those relating 
to the flint lens by the subscript 2. 

Since i ^ /i i\ 

if /&, changes to /a, + A/a„ the corresponding change produced 



I 
m -7 IS 



and in the same way the alteration of j due to an alteration 
A^ of ^ is _ V. i_ 

u is the distance of the point at which the Hp rays and the 
H rays are united by the lens, and hence is the same for both. 
The total effect of the change in the refractive indices is there- 



A PHOTOGRAPHIC CORRECTING LENS IO9 

If the symbols in this formula relate to the Hfi line, the 
change of v, to satisfy the conditions for the proposed lens, must 
be equal to the difference between the ordinates at //S and Hfi in 
the color curve of the objective. 
From the table of data we take 

/A, = 1. 5231, ^1^ = .0078, V = i.oooo for the Hfi line. 
/i^ = 1.6347, Afi^ = .0171, V = 1.0701 for the Hi line. 

Also A-= —0.06551, and substituting these values the condi- 
tion becomes .014911 , .026942 

—^;—+—^F— = -. 06551. 

If we wish to determine the lens so that the position of the 
focus for the Hfi line shall remain unchanged, then for these rays 
the lens must act like a thin glass plate, and we have the further 
condition /, = — yi, 

and combining this with the preceding one, 
OMQ" .026942 _ 

-7 ;^--<^655i, 

from which >. , ^ , 

/, = -f- 0.18365 meters 

/.= -0.18365 " 

If we make the flint lens double convex with equal radii, and 
the inner surface of the crown lens to fit the flint, the radius of 
each of these surfaces will be 0.233 meters, and the radius of the 
back concave surface of the crown lens 0.163 meters. Such a 
lens would be perfectly easy to make. The surfaces in contact 
could be cemented to diminish loss of light by reflection, and a 
lighter flint could be used than that chosen for illustration. With 
the flint lens turned toward the objective, the lens would, more- 
over, have very little spherical aberration, although determined 
by these approximate formulae. 

In order to find the form of the color curve when this lens is 
placed in position, we first require the focal lengths of each lens 
of the combination for different rays of the spectrum, — or rather 
the reciprocals of the focal lengths, which are most conveniently 
obtained by means of the relation 



no 



JAMES E. KEELER 



I 
— 77 



/I— I 



where /*',/', are respectively the refractive index and focal 
length for the Hfi line, and /*, / the corresponding values for any 
other line. For any one line of the spectrum we then substitute 
the values so found, together with the value of v for the same 
line, as given in the table of data, in the equation 

I _ I I I 

and find u, which is the distance at which the same rays are 
united after passing through the lens. The following table con- 
tains the results : 



Line 


V 


\ 


I 


I 


I 


u 






V 


fx 


/. 


u 




B 


I. 0000 


I. 0000 


53464 


—5.2821 


0.9357 


1.0687 


H^ 


0.9939 


I. 0061 


5.3567 


-5.2976 


0.9470 


1.0560 


D. 


0.9886 


I.0II5 


5.3827 


-5.3396 


0.9684 


1.0326 


E 


0.9905 


1.0096 


5.4160 


-5.3954 


0.9890 


I.OIII 


^ 


0.9914 


1.0087 


5.4234 


-5.4065 


0.9918 


1.0083 


m 


I. 0000 


I. 0000 


5.4452 


-5.4452 


I. 0000 


I. 0000 


Hy 


1.0368 


0.9645 


5.4962 


-5.5361 


1.0044 


0.9956 


m 


I. 0701 


0.9345 


5.5264 


-5.5919 


I. 0000 


I. 0000 



From the last column of this table, the color curve shown in 
Fig. 3 has been platted. It is obviously much better adapted to 
photographic work than the original curve. With the spectro- 
scope slit in the focal plane of the Hy rays, the //)8 and Hi rays 
would be only 4""°.4 out of focus. The spectrum would be 
practically linear for a considerable distance on each side of //y. 

In practice, the lens could be mounted in a cell provided 
with suitable adjusting screws, on a swinging arm within the 
telescope tube, so that it could be pushed into place by a rod 
projecting through the tube, and held centrally by adjustable 
stops. When the telescope was required for ordinary visual 
observation the lens could be withdrawn from the cone of rays. 
In the future a lens of this kind will, no doubt, be regarded as a 



A PHOTOGRAPHIC CORRECTING LENS 1 1 1 

necessary adjunct to every large refractor. Of course it would 
be easy to make a series of such lenses (three would probably 
be sufficient) by which the spectrum could be made linear at a 
number of different important points. 

The only difficulty connected with the method is that the 
correcting lenses could not be made until after the color curve 
of the objective had been determined ; i. ^., until after the erec- 
tion and adjustment of the telescope. This is, however, not a 
very serious matter. 



SCHMIDTS THEORY OF THE SUN. 

By E. J. WlLCZYNSKI.« 

Considering the rapid progress which has been made in the 
observational or practical side of solar physics, it must be con- 
fessed that the theoretical side has been very imperfectly devel- 
oped. Almost every student of solar physics has his own 
theory, and usually he himself is the only one that believes in it. 
Under such circumstances we ought to welcome a theory which 
has indeed some difficulties to overcome, but which is neverthe- 
less founded on physical laws that are accurately known, and 
which is capable of an exact mathematical treatment. Never- 
theless, the theory of Schmidt, the fundamental principles of 
which are explained in a small pamphlet, " Die Strahlenbrechung 
auf der Sonne; ein geometrisches Beitrag zur Sonnenphysik," 
von August Schmidt, Stuttgart, 1891, has, even here in Ger- 
many, not met with full appreciation. In spite of its impor- 
tance, I know of no English article upon the subject, and I will, 
therefore, present the most important points in this paper. The 
theory, it is to be noted, does not undertake to explain every- 
thing which takes place upon the Sun. Certain fundamental 
facts are explained by it, and all others are left to be explained 
by auxiliary theories, which will be of the character of the solar 
theories already current. 

The theory is based upon the ordinary laws of refraction. If 
we regard a ray of light which is horizontal at a certain point 
of the Earth's surface, we shall find by the formulae for atmos- 
pheric refraction that its curvature is about one-seventh of the 
curvature of the Earth's surface. Now let us suppose that we 
have a body whose radius is seven times that of the Earth, while 
the composition of its atmosphere, as well as the conditions of 
temperature and pressure at its surface, are the same as at the 
Earth's surface. The ray then has the same curvature as the 

' Student of Astronomy, University of Berlin. 



SCHMIDTS THEORY OF THE SUN 1 13 

surface of the body. It will therefore describe a circle, and make 
the entire circuit of the body. If the body is still larger, or if 
the density of the atmosphere (or more properly its refractive 
power) is greater, a ray which is horizontal at a given point of 
the body's surface will be unable to pursue its course out into 
space, but it will return to the surface. The ray will require a 
certain angle of elevation in order to leave the body. At a 
certain height above the surface the path of a ray will be circu- 
lar, and above that the refraction will be the same as in our 
terrestrial atmosphere. 

The investigation of this general case of refraction was first 
made by Kummer,' and his most important results are given 

below. The velocity "of light at any point of its path is «' = y 

According to the undulatory theory, the time which the light 
requires to pass between two points A and ^ is a minimum. 
That is, ^B 

A 

must be a minimum. Therefore 






:0. 
V 
A 

The velocity v is inversely proportional to the index of refrac- 
tion fi of the medium. Hence 

B B 

sCfuis=Ci(fjLds) = o, 

A A 

Performing the variation, which offers no difficulty, we have 

wt«'+l«^+^)]^yf{|-l(4)}«' 

+{fM(4)>«.'+{t-|(''s»H*=°- 

A and B are fixed points, for which &r, 5y, 8« are, therefore, equal 
to O. This condition is fulfilled only if 

* SUmngsbiricHii der Berliner Akademie, 12 Juli, i860. "Ueber atmosphiirische 
Strablenbrechang." 



114 E. /. WILCZYNSKI 

d I dxK du, d i dy\ du, d t dz\ dii . v 

These are the differential equations of the ray. Only two of 
these equations are independent. The third can be derived 
from the other two, as may be easily shown. Hence two of 
them suffice to determine the path of the ray. 

We now proceed to specialize the equations ( i ) for the case 
which is alone of importance to us, viz.^ an atmosphere, surround- 
ing a spherical body, whose index of refraction decreases with the 
altitude above the surface but is the same at all places of equal 
height. In this case /ui is dependent only upon r=i/jc*+y + «'. 
The path of the light is obviously a plane curve. Hence we may 
choose its plane as the plane of xy^ or put s=^o. We have, 
therefore, the equations 



^/ dx\ djA di dy\ d^k 

Hy^Hflx 'ds\^dsf~'^ 



to determine the ray. Since iL^f{r)^ and x and y occur only 

in the combination r, we have : 

d^i ^\ dfA dfkx dt ^\ dj^ dy^ y 

Hy^lsf'dx^'d^ r 'ds\^'ds^^'dy^~^y 

We easily find after a simple transformation, 

and integrating. ^(,|__,g)^c: (2) 

dx dv 

Now we have — = costf, ^ = sintf, if 6 is the angle between the 

tangent to the curve at the point x^y and the positive axis of x. 
(Fig. I ) In polar codrdinates we have 

x = r cos ^, j^ = r sin ^, 
hence according to (2), and putting $ — ^=180^ — a, 

/ir sin (0 — ^) = ^sin o = C. (3) 

This well-known relation, which can be very simply derived in 
another elementary manner, is very important for our purpose. 
Schmidt bases his entire theory upon it. 



SCHMIDTS THEORY OF THE SUN 1 1 5 

We can integrate (2) by introducing polar coordinates, r 
and ^. (2) becomes 

Cdr 



According to (3), if the ray starts from the surface of the body 
whose radius is R^ and where i^=^i^w and = 90^ — 1, we have 

flo-^ cosi= C 
If we put r=-^+A, where h is the height above the surface, and 
if the ray starts from the point A=o, ^=:o, we have finally 

IK^R cos I dh 
{R + A)i/^-(^ + ^)--/*o'^ cos-i 



J (R ' ^^-/-■/» ■ ^^" . D. — • -• ' w 




Fig. I 

This is the integral given by Kummer. The above rather more 
detailed investigation is taken from the still more explicit Habili- 
taiionsschrift by Dr. O. Knopf,' to which we shall frequently refer. 

We here distinguish three cases. 

I. For all values of h and 1, /*(-^ + ^)>fK^-^ cosi. cosi has 
its maximum value for / = o. Hence our condition is 

In this case there is a real, finite value of ^ corresponding to 
every real and finite value of h. In other words, in this case a 
ray which is horizontal at the surface of the body passes out 
into space. Its curvature is less than that of the surface of the 
body. This is the ordinary case of refraction in the terrestrial 
atmosphere. We shall call bodies of this kind, bodies of the 
first class. (Fig. 2) 

' Dii Sekmidfsche Sonnentheorie und ihn Anwendung auf die Methode der spekinh 
skopiuken Besiimmung der Rotaiionsdauer der Sonne, Jena, 1893. 



1 16 E, J. WILCZYNSKI 

2. It can happen that, for a certain value of h. 

According to (3) this means that a ray making the angle $ with 
the horizontal line at the surface, after refraction reaches the 
height k and is there horizontal. But in this case we have ^ =qo , 
as may be proven by applying the following theorem of the 
integral calculus : 

It A^) becomes infinite for jc=Jf, and we can find a value 
o between Xo and X such that {x — XY/{x) or {X—xY/{x), for 
ff > I, is greater than a certain number K for all values of x 

between a and X, then the integral j/(x)dx becomes infinite. 




Fig. 2 

Hence a ray tangential to the sphere at this elevation makes 
an infinite number of revolutions around the circle whose radius 
is ^-f^. We have here found the analytical condition of circu- 
lar refraction. 

3. For certain values of A and i we may have 

For these values of A and 1 the ray cannot exist, since ^ becomes 
imaginary. We are especially interested in the ray which leaves 
the surface horizontally, i.e,, for which 1 = 0. Our condition is 
then fjL{J^ + AXfi^/i. 

Let us now examine the remarkable appearances which 
would present themselves to an observer on a body whose 
atmosphere fulfils the second and third conditions. (Fig. 3) 
We call such a body a body of the second class, while those 
bodies whose atmospheres fulfil the first condition only will be 



SCHMIDTS THEORY OF THE SUN II7 

called bodies of the first class. It may be said, by the way, that 
Jupiter is probably a body of the second class. 

If an observer on such a body directs his eye horizontally, 
he will not, as on a body of the first class, see the horizon. 
Moreover, a ray which reaches his eye horizontally will have 
come from another point of the body's surface, which it left at 
a certain angle of elevation. The same holds true if he grad- 
ually directs his eye upward until he reaches a certain angle of 

elevation i, given by cosi = ^^^^ — ^^t_J. At this elevation the 




Fig. 3 

first ray coming from outside of the body can reach his eye, and 
this is his horizon, which has, as we see, the elevation i. The 
observer will therefore have the impression of standing in a 
hollow shell. As he raises his eye from 1=0 to 1 he will succes- 
sively see more and more distant points of the surface; at 
length he will see his antipodes, and, raising his eyes still more, 
his own back. Still raising his eyes, he will again see all the 
points he saw before, and so on for an infinite number of times. 
Of course all of these images will be distorted and partly over- 
lapping. If the observer looks at the zenith and lets his eye 
sink he will gradually see the entire sky, from the zenith to nadir, 
and again from nadir to zenith, etc., where of course the images 
are again distorted. That the observer sees an infinite number 
of images of the same object in this manner can be analytically 
shown. We note that outside of the sphere of circular refrac- 
tion the refraction is of the ordinary kind. An observer placed 



Il8 E. J. WILCZYNSKI 

there would see none of the above phenomena. The rays 
which, coming from space, become nearly tangential to the 
sphere, on piercing it at B reach the surface under an angle 
which differs only slightly from i, at a point A, Let us change 

the direction of the eye at A by di. If we find the value of -j. 

for the value of h corresponding to the sphere of circular refrac- 
tion, we shall know the change in the position of B correspond- 
ing to the change of di in i. This is equivalent to knowing the 




Fig. 4 

change in the direction of the point observed corresponding to a 
minute elevation of the line of sight. We find : 



(s) 



^ ^ /* V*. d_ 1- cos/ -.,, 

- ~'^^ "'"y [M'(iP+>i)'-M.'^ cos-/]* • 

o 

For the sphere of circular refraction we have : 

d^ 
By applying our criterion, we find that for this value -r. becomes 

infinite, i, ^., to a small change in i corresponds a very large 
change in ^ (Fig. 4), in fact an infinite change in ^. We there- 
fore see an infinite number of nearly coincident images of the 
same object. By seeing is meant, that the theoretical course of 
the rays is such as to enter the eye. As a matter of fact, the 



SCHMIDTS THEORY OF THE SUN II9 

absorption will probably prevent even all of the first images from 
being seen. 

We have now reached the most important part of our paper, 
which treats of the appearance that such a body of the second 
class presents to an outside observer, whose distance from the 
body's center is e. The observer is supposed to be outside of 
the atmosphere. Then we have, for the observer, r = ^, fi= i, 
a = y, if y is the apparent angular distance from the body's cen- 
ter of that point from which the ray issues. Then we have : 

y^^R sin a = ^ sin y. 
Let y. be the real angular distance, as seen without an atmos- 
phere, corresponding to the apparent y; then we have very 
nearly .^ = irsiny. 

if we consider only the case that y and y, are small angles, or 

that e is very large. Under the same supposition, we have for 

the magpiification of the angular distance of these two points, 

y sin y 

— =-: — ^ = A*o sma, 

y. smy, '^^^ • 

In bodies of the first class a can assume the value — . In bodies 

2 

of the second class a will always be less than a certain angle ^ 

less than — . All rays making an angle greater than fi with the 

normal at the body's surface, will not find their way into the 
observer's eye. Hence the magnification of R in this case is 
fi^sin ^. 

Let us now follow the ray which, rising from the surface 
under the angle of elevation 1, reaches the sphere of circular 
refraction r. horizontally. This ray will describe the circle whose 
radius is r. an infinite number of times, and as has been shown, 
cuts it at last to pursue its path out into space. The ray which 
diCEers from this only slightly will seem to the observer to come 
from a point whose distance from the center differs slightly from 
mr.. In other words, while the real diameter of the body is R^ 
it will seem to be p^r^. 

We will now consider the case in which the radius of the 



120 E. /. WILCZYNSKI 

solid body ^ = o, i. ^., the case of a gaseous body made up of 
spherical shells whose index of refraction decreases as the dis- 
tance from the center increases. We will moreover assume the 
gases to be in a state of incandescence, and will investigate the 
appearance which such a body will present to an outside 
observer. 

In a large body the pressure, and hence the density, of the 
gases in the center will be very great. If the temperature is 
above the critical temperature of the gases, these will not be 
condensed. Nevertheless the density will be so great that the 
gases will emit white light. Let the critical sphere of circular 
refraction have the radius r„ then we have 

^r sin o = ^ r. 

for a ray which leaves this sphere tangentially. This ray is, 
however, also tangential to a smaller sphere r^ for which 

This ray after leaving the sphere r^ tangentially, cuts the suc- 
cessive spheres r under the angle f, g^ven by 

COS I == sm a = ^-^ 

f^ 

It then becomes tangent to the sphere r,, and after making an 
infinite number of revolutions around the circle whose radius is 
r., leaves this sphere and proceeds on its path. This ray can, of 
course, only reach the observer after an infinite time, and after 
experiencing an infinite amount of absorption, i. ^., not at all. 
But those rays which rise from r^ under a very small angle of 
elevation will have a finite path, and will reach the eye almost in 
the direction of a ray issuing tangentially from r,. Moreover, 
no rays tangent to the spheres r between r^ and r, can pass into 
the observer's eye ; for if there were such a ray we should have 
for To < r < r;, fir = ^r^ sin Oo = ^r, sin 0^, hence ^ = 00 . That 
is, between r^ and r, only the bounding spheres r^ and r, can send 
tangential rays into the observer's eye. (Fig. 5) It follows 
from this that in the image of the body seen by the observer, 
the space between r^ and r, is not represented by a correspond- 



SCHMIDTS THEORY OF THE SUN 121 

ing ring. Moreover the spheres r^ and r, will seem to coincide, 
since the ray tangent to r^ is also tangent to r,. If r, = ^ were 
the radius of the solid body which the atmosphere surrounds, it 
would appear to the observer as if the body had the radius r.. 
The height of the body's atmosphere would be shortened by the 
quantity r, — r.. If, however, as we here suppose, the inner 
spheres consist of incandescent gases above their critical tem- 
perature and emitting white light, while the outer spheres consist 
of incandescent g^ases with a gaseous spectrum, whose light, 
owing to the smaller pressure, is less intense than that of the 
inner spheres, the observer will notice a sudden change in the 




Fig. 5 
intensity of the light. For the image of the sphere r^ is imme- 
diately bounded on the outer side by the image of the sphere r., 
whose light is much less intense. The difference of intensity 
will obviously depend upon the distance r, — r^. 

If we consider the Sun to be an incandescent gaseous sphere 
of the second class we here evidently have an easy and perfectly 
natural explanation of the sharp apparent boundary of the Sun 
called the photosphere. The theory which considers the Sun in 
the main as liquid, has been abandoned by most astronomers. 
According to it the photosphere found a comparatively reason- 
able explanation, as the surface of the liquid upon which the 
atmosphere rests. The theory which has been most generally 
accepted regards the Sun as in general gaseous, and the (iphoto- 



122 E. J, WILCZYNSKI 

sphere as a layer of clouds containing small particles of con- 
densed liquid substances. Now the lack of any dispersion on 
the Sun's limb, which is also explained by our theory, shows 
that the atmosphere of the Sun directly above these hypothetical 
clouds is extremely tenuous. It is true that our terrestrial clouds 
are also supported, perhaps by air currents, at altitudes which 
they could not reach according to the law of specific gravity 
alone. But these solar clouds are very much heavier than the 
terrestrial, and the gases in which they float are very much less 
dense than the terrestrial atmosphere at the altitude of our clouds. 
The extreme uniformity of their height also renders this theory 
improbable. Moreover, I think that we are justified in consider- 
ing that the temperature of even the outer parts of the Sun is 
above the critical temperature of the metallic vapors. 

If therefore we accept the very plausible theory that the 
Sun is nothing but a huge incandescent gaseous ball, we arrive 
at the result that to our eye this body would present the same 
appearance which the Sun actually presents ; t/tsr., a sudden 
change of brightness at a certain distance from the center ; that 
is, we explain the distance of the photosphere. 

This is the essential part of the theory. Obviously the 
explanations of Sun-spots, etc., hitherto accepted, are not at all 
affected by it. On the contrary the possibility of explaining 
them is increased, inasmuch as we can suppose their real situa- 
tion to be at any point of the gaseous ball, and we are not 
restricted to the surface. It must here be noted that any attempt 
to explain the prominences as simply an effect of irregular refrac- 
tion, as Schmidt does, is altogether unauthorized by our present 
physical theories. Such irregular refraction cannot alter the 
wave-length of the light. As long as we have no evidence that 
the change in the wave-length of a certain line, as shown by the 
spectroscope, may be produced by other causes than motion in 
the line of sight, we have no right to try to explain away veloci- 
ties which we think are too great to be credible. The spec- 
troscope says that these velocities are actually present, and 
no thoory, however ingenious, can ignore this fact. For the 



SCHMIDTS THEORY OF THE SUN 1 23 

same reason, such theories as that of Brester are at present 
unjustified. 

It has been established by spectroheliograms taken by Pro- 
fessor Hale, and by corresponding visual observations, that the 
prominences as shown in the C line, and in the K line of calcium, 
are very similar. It has sometimes even been found that the 
calcium image was higher than the hydrogen. But it seems 
very difficult to suppose that in the Sun calcium vapor should 
be found in the same or in a higher layer than hydrogen. From 
our standpoint we might suppose that the actual origin of the 
prominence is in the calcium layer below the hydrogen, and 
that this eruption, if we so call it, disturbs the overlying layers 
of hydrogen, and causes them to assume a form similar to that 
of the calcium prominence. Now, according to our theory, 
this second prominence could be placed in the path of the 
rays passing from the first into the observer's eye. The two 
would then appear superposed. The objection might be raised 
that according to the law of diffusion of gases we should not 
expect the gases to be, as it were, sorted into layers, according 
to their density. Still, observation seems to speak for this 
arrangement, at least in a general way. In the prominences we 
have hydrogen and helium; still higher we have coronium. 
Occasionally the prominences contain sodium, magnesium, etc. 
All this seems to point to a stratified condition of the solar 
atmosphere. It might, however, be reasonably expected that 
not only calcium, but the metals sodium, magnesium, etc., with 
an atomic weight smaller than that of calcium, would be regu- 
larly present in the prominences, if this explanation is correct. 
There is certainly some evidence for this, and it is perhaps worth 
while to make special investigations as to the presence of these 
metals in the prominences. Possibly their lines sometimes 
escape notice through insufficient brightness. 

At any rate it is interesting to note how, in this way, many 
observed facts are explained. 

It has been remarked by Seeliger that the ray could never 
reach the observer's eye after having followed such a long path. 



124 E. J. WILCZYNSKI 

owing to the absorption. But it must be noted that when a white 
ray passes through the several incandescent gaseous strata, these 
cause only selective absorption, and leave unweakened those parts 
of the spectrum whose wave-length is different from that of the 
light which the gases emit. We have here also a natural explana- 
tion of the origin of the Fraunhofer lines, which does away with 
the strained and factitious theory of a thin reversing layer cover- 
ing the photosphere. It is certainly very unreasonable to assume 
that all of the substances whose presence is indicated by the 
Fraunhofer lines should be collected in such an extremely thin 
layer. 

The small mean density of the Sun is explained by our theory. 
The atmosphere of the Sun is exceedingly tenuous above the 
apparent limb, and the great difference in the intensity which is 
found there seems to indicate that the origin of the white light, 
the real photosphere, is at a considerable distance below. This 
central part must have a great density. We can easily explain 
the absence of dispersion at the Sun's limb by the small density 
of the atmosphere above the sphere of circular refraction. We 
there have, if r, is the radius of this sphere, m, = m,'> if fi, and ft/ 
are the indices of refraction for red and violet rays respectively. 
Hence /*,r, = /*/ r/ since fi^ = ^/ and r^^rj . We have, more- 
over, ff»^o = M,^,. andfio''*©' =/*,'^,'. Hence f«toro = /i^Vo'. Butin 
the sphere whose radius is r^ the density will be considerable, 
and we shall have fK^'>/io- Therefore rj<jr„ i, e,, the violet rays 
which are tangent to the sphere r, come from a lower sphere 
than the red rays. The two together give white light again. 
However, on analyzing this more closely it will be found that 
the violet light, having pursued a longer path, will be' fainter 
than the red, even if we suppose the solar atmosphere to act 
upon red and violet light in the same way. Between r, and r,» 
and rj and r/ =r, the effect of dispersion will be to unite the 
violet rays from a lower sphere with the red from a higher; 
hence there will be no dispersion on the limb. But the intensity 
of the violet rays will, upon analysis, be found to decrease more 
rapidly than that of the red as we pass from the center to the 



SCHMIDT'S THEORY OF THE SUN 1 25 

limb, and this is in accordance with observation. However, this 
latter fact is explained in an equally satisfactory manner by the 
old theory. 

An application of Schmidt's theory to the spectroscopic 
method of determining the period of rotation of the Sun, is given 
by Dr. Knopf. As we have noted, the place where absorption 
causes a certain line in the spectrum is not at the apparent sur- 
face, but deeper within the Sun. This place of absorption may 
be represented by A, If we suppose the ray to be emitted from 
A^ where the radius is r, under the angle a, and to leave the 
sphere r, tangentially at B^ it may be proven that all rays leav- 
ing the sphere r, under an angle which differs very slightly from o, 
will leave the sphere r, at B very nearly tangenHally. Now the 
slit of the spectroscope has a certain width. It will hence 
receive all rays issuing from B, a point on the limb, which make 
an angle smaller than a certain maximum value with the hori- 
zontal line at B, The observed displacement of the line will 
then be the mean of the displacements corresponding to the 
velocities of all points which send rays into the slit» 1. ^., of all 
points on the entire circle whose radius is r, and whose plane is 
determined by the Sun's center, the point B in which they leave 
the Sun's limb, and the observer. Knopf did not plainly say in 
his paper that he did not mean one single ray, but an entire 
bundle, or that this consideration of more than one ray is made 
necessary by the appreciable width of the slit, and he did not 
prove that to a very small difference di of the angle 1 which the 
ray makes with the tangent to the sphere r,, corresponds a very 
large change in the position of the point A in which the ray 
pierces the sphere r. This latter proof we have given in proving 
that for r = r. d^ 

which we deduced from (5). We there considered A as fixed, 
and varying / by di\ obtained d^, the change in position of B. If 
we consider B as fixed, as we do here, d^ gives the change in 
position of A^ the opposite sign obtained in that case being of 
no importance. 



126 E. /. WILCZYNSKI 

It is evident that the component in the line of sight of the 
velocity which a point of the circle r has in its own plane is 
alone of importance in producing a displacement of the line. 
For this component Knopf finds the value 

sing cos » 

cos^ -^^'^'^^' 

where ^ is the heliographic latitude of the observed point of the 
limb B, and ^ is the heliographic latitude of that point A inside 
the Sun's apparent surface, at which the absorption causing 
the line takes place. The point A may be determined by the 
angle v which OA makes with OB. The mean of the velocities 
of all points A of the circle r, will be obtained by integrating the 
above expression from v = o to v = 2w and dividing by 2w, u e.^ 



if' 



sin a cos^/(r,^)^ 
2w/ cos tjf 



If the absorption-line is caused by a stratum between r' and r' 
the mean of all rays from this stratum will have to be formed. 
This is r- »» 

cos ^ / / sin af{r^^^drd¥ 
2»(r'— r')/ / Tl^ ' 



where /(r, ^) is the law according to which the velocity of a point 
of the Sun depends upon its latitude and distance from the 
center. In deducing this expression it has been assumed that the 
Sun's axis of rotation is at right angles to the line of sight, which 
takes place twice a year. Knopf also treats the more general 
case where the inclination is any angle, and from the formulae 
which he obtains, arrives at the conclusion that if Dun^r's values 
are correct it is impossible that the Sun rotates as a solid body, 
or that the concentric shells rotate as solid shells. For the 
details the reader is referred to Knopf's paper in A. N. 3199. 

In closing this paper I may express the hope that American 
astronomers will recognize the elegance and beauty of Schmidt's 
theory, and will contribute to its further development. 



A CLOUD-LIKE SPOT ON THE TERMINATOR OF 

MARS. 

By A. E. Douglass. 

On November 25 and 26 a bright spot was seen in the unil- 
luminated portion of Mars, to which, in my opinion, no other 
name than cloud can be applied. Its great height, size, and 
brilliancy, and, on the second evening, its singular fluctuations, 
render it of importance in the study of the Martian atmosphere. 

I first saw it at le*" 35", G. M. T. of November 25, and made 
an estimate of its height. It seemed to be rapidly increasing in 
length in a direction parallel to the terminator at that point. Sub- 
sequent estimates of its height gave a different and greater value 
than at first, until its sudden disappearance at 17^ d'^or perhaps a 
minute later. After once attaining its size, it seemed to remain with 
little change, presenting the appearance of a line 140 miles long 
by 40 miles wide at the center and lying parallel to the terminator, 
but separated from it by an apparent space of over 100 miles. 
It was generally yellowish in color, like the limb, but of less bril- 
liancy than the center of the disk, though distinctly surpassing 
in that respect the adjacent terminator. I estimated it to have 
the brilliancy of the light areas of the disk at a distance of 9° 
from the terminator. In one view it appeared to be a very small 
whitish point (Observation 2, below), and I am inclined to think 
that there may have been a real diminution in its size at that 
moment. This idea is partly sustained by the following night's 
observations. At 16*" 54" it was observed by Professor Pickering, 
whose estimate of its height is found in Observation 6, below. 
At 17*" 5", after obtaining two readings of the micrometer screw 
for latitude, the seeing, which had been quite steadily at the fig- 
ure 7 (on a scale of 10), dropped to 4, and in attempting the next 
setting I could not find the "cloud," although once before it had 
remained visible when the seeing dropped instantaneously to 
that figure. Nor did it reappear in the next half hour. This 

127 



128 A, E, DOUGLASS 

sudden disappearance without any previous lessening of its 
height above the terminator or of its size, made its cloud char- 
acter unmistakable, since a mountain beyond the sunrise termi- 
nator must either constantly decrease in height, or soon join to 
the illuminated disk. 

A subsequent computation showed that this phenomenon 
took place over the southern part of Schiaparelli's Protei Regio. 
Other reasons lead me to think, however, that he has placed 
that island some 5^ too far south. 

On November 26 the cloud promptly appeared at 17^*1 5" 
G. M. T., but nearly 9° farther north. Instead of remaining 
continuously visible it dissipated and re-formed at irregular 
intervals. The first appearance lasted sixteen minutes. After 
somewhat over four minutes had passed it reappeared momen- 
tarily, and six minutes elapsed before it appeared again, last- 
ing then but two and one-half minutes. Then followed an 
absence of three minutes, presence for two minutes, absence for 
three minutes, presence one minute, and a final brief appearance 
eight minutes later at 18^ i"". Its presence was suspected five 
minutes before that hour, and again at 18^1 1"*, but with great 
uncertainty. 

At this time it presented in general the same characteristics 
as the night before, though its appearances were too brief to 
permit such careful observations as were hoped for. The seeing, 
too, was not so good as before, varying from 4 to 7, and if the 
cloud happened to appear under the former figure, its observation 
was difficult. It is needless to remark that under such condi- 
tions it was impossible to observe its appearance or disappear- 
ance to the second. In general, it seemed to exhibit a less eleva- 
tion than the night before. A careful estimate of its latitude 
placed it precisely at the center of the terminator. I believe 
these latitude observations, though made rapidly, cannot be sub- 
ject to an error greater than 2^, and probably less than i^. On 
November 27 at 18^ I searched for the cloud, but was not 
rewarded by finding any trace of it. 

Estimates of the size and height of this cloud were made 



A CLOUD-LIKE SPOT ON MARS 



129 



with reference to a glass thread in the micrometer, whose diam- 
eter is o'.Q. One-tenth of the thread, therefore, represented on 
Mars a little less than twenty-four miles. In Table I below the 
observed dimensions are given directly in miles. Column S gives 
the separation of the cloud from the terminator ; W gives its 
width ; H, its total height, and L, its length along the termi- 
nator. In the two final columns the longitude and latitude of 
the nearest point of the terminator are given. Table II gives 
various computed quantities. 







TABLE 


I. 










Date 


Obwr. 


Time (CM. T.) 


s 


W 


H 


L 


LoDf.Term. 


Lat. 






h m 














Nov. 25 


I 


16 37.5 


71 


4» 


119 




43.6 


-32.5 


M 


2 


39.0 








142 


44.0 


t« 


U 


3 


4I± 


143 




143 




44.4 


<( 


" 


4 


47.0 


119 


5« 


177 




45.9 


4( 


M 


5 


49.1 






166 


142 


46.3 


« 


M 


6 


55.0 


119 








47.9 


<• 


Nov. 26 


7 


17 23.0 


96 


24 


120 




43-0 


-23.3 


(« 


8 


43.0 






118 




47.8 


M 


M 


9 


49.3 


49 


48 


97 




49.4 


M 



TABLE IL 



Obwr. 


Min. Ht. 
Bottom 


Loag. 


Diat. Term. 


Vcrt.Ht. 
Top 


Min. Hl 
Top 


Lomr. 


Dist. Tenn. 


Lat. 


I 
2 


6 


48.7 


159 


58 


17 


51.9 


263 


-30.3 


3 


24 


54.5 


323 


24 


24 


54.5 


323 


<t 


4 


17 


54.2 


263 


79 


38 


58.2 


399 


«( 


5 










34 


58.0 


376 


" 


6 


17 


56.2 


263 










(4 


7 


II 


49.1 


211 


37 


16 


50.5 


263 


-21.5 


8 










16 


55.2 


259 


« 


9 


3 


52.6 


107 


56 


II 


55.5 


211 


«4 



In Table II, column i gives the number of the observation, 
column 2, the height at which the shadow of the planet will 
strike a cloud presenting the observed separation ; column 3 
gives the surface longitude for such a point ; column 4 gives the 
tangential distance in miles of such a point from the terminator ; 



130 A. £. DOC/GLASS 

column 5 gives the total height the cloud would have if all its 
apparent height were extended vertically over this point ; column 
6 gives the least possible height the top could have and still be 
illuminated by the Sun ; columns 7 and 8 give the longitude of 
this point and distance from the terminator, and column 9 gives 
the latitude. 

In order to get an idea of the mean height of this cloud, we 
may take the mean of column 4 and average it with the mean of 
column 8, obtaining 260 miles. This gives us an elevation above 
the surface of between 16 and 17 miles. In this process we 
have taken the apparent center of the cloud, and have assumed 
the seeing to have no influence. We obtain, therefore, the 
smallest possible mean height of the center of the cloud. If we 
assume that the seeing was not perfect, its effect would be to 
lessen the separation, but not to change the total height. Sup- 
posing, for example, that the apparent extension of the cloud 
was due to poor seeing enlarging a point, then our terminator 
distance would be 299 miles, and our minimum elevation 22 miles. 
Therefore we can assume 20 miles to be the smallest probable 
mean elevation of this cloud. The average height of our cirrus 
clouds is five and one-half miles. 

One more idea requires mention, namely, the movement of 

this cloud in latitude. From the extreme rarity of such an 

occurrence I am inclined to connect intimately the appearances 

of the two evenings, and consider them as due to one source, 

presumably a large body of air moving northward. Such an 

advance would be at the rate of 13.1 miles per hour. 

Lowell Observatory, 
December 10, 1894. 



PRELIMINARY TABLE OF SOLAR SPECTRUM 
WAVE-LENGTHS. II. 





£ 


\y Henry A 


. Rowland. 










Intensity 






Intensity 


Wm-leacili 


Subttanoe 


and 


Wsve*wngtii 


Subttanoe 


and 
Character 


3911.554 


Mn 


oNd? 


39x7.400 






39II.708 




000 


39X7.73X 


Cr 




3911.836 


Fc 


I 


39x7.893 




000 


3911.963 


Sc 


2 


39x8.007 




oNd? 


3912.127 


Cr? 


2N 


39x8.223 




ON 


3912.224 







39x8.396 


Mn 




39x2.341 


V? 





39x8.464 


Fe 




3912.445 


Ni? 


2 


39x8.563 


Fe 




39x2.561 







39x8.713 






39x2.732 




00 


39x8.789 


Fe 




3912.935 


Fc 





39x8.929 




00 


3913.030 




00 


39x9.035 






39x3.123 


Ni 


2 


3919.208 


Fe 




3913.282 




oooN 


39x9.309 


Cr 




3913.395 




1 


39x9.499 




oooN 


3913.609 


Ti-Fe 


5d? 


39x9.708 




oNd? 


39x3.775 


Fe 


4 


39x9.869 






39x4.153 


Ce? 


iN 


39x9.956 


Cr 




39x4.320 







3920.x X4 






39x4.426 


Fe? 


3 


3920.264 


Co? 




39x4.477 


Ti 


2 


3920.410 


Fe 


10 


3914.566 




I 


3920.59X 




oN? 


39x4.652 


Ni? 


I 


3920.768 


Fe 




39x4.880 


Fe 


oN 


3920.868 


Co 


IN 


3915.094 




oN 


3920.984 


Fe 




39x5.359 


Fe 


I 


392X.105 






39x5.612 


Fe.Cr 


I 


3921.188 


Cr-Nd 




39x5.751 




3 


3921.326 






39x5.951 


Cr- 


5d? 


392X.4X5 


Fe 




3916.079 


Zr 


I 


3921.563 


Ti 




3916.207 


ZrLa 


oNd? 


3921.69s 


La- 




39x6.383 


Cr 


2 


3921.855 


Zr-Mn 




39x6.54s 




3 


3922.043 




oNd? 


3916.661 


Mn 


00 


3922.159 




00 


39x6.745 


Zr 


00 


3922.223 


Mn 




39x6.879 s 


Fe 


5 


3922.560 


V 


IN 


39x6.992 




00 


3922.81S 


Mn 




39X7.X25 







3922.907 


Co 




39x7.264 


Co 


2 


3923.054 


Fe 


12 d? 


3917.324 


Fe 


5 


3923.X80 










i: 


JX 







132 



HENRY A. ROWLAND 







Intensity 






Imenaity 


Wave-length 


Subttance 


and 
Character 


Wavelength 


Subatanoe 


and 

Chaiacter 


3923.246 







3930.654 







3923.375 


Mn 


00 


3930.804 







3923.472 


Sc 





3931.030 







3923.641 







3931.269 


Fc 


I 


3923.831 




ooN 


3931.483 




ON 


3924.065 




000 


3931.729 


Mn?- 


IN 


3924.206 


Mn 




3932.039 




ON 


3924.313 






3932.161 


Ti 


I 


3924.492 




ooN 


3932.395 







3924.673 s 


Ti 




3932.625 




ooN 


3924.791 






3932.785 


Fc 


I 


3924.929 




00 


3933.056 


Fc 





3925.153 




ooN 


3933.523 




8N 


3925.347* 


Co, Fc? 




3933.825 S K 


Ca 


< 1000 


3925.491 




000 


3934.108 


Co- 


8N 


3925.677 






3934.174 




oN 


3925.790 s 


Fe 




3934.661 




oN 


3925.939 




000 


3934.818 


Co ?, Fc ? 





3926.086 ^ 


Fe 


f^ 


3935.358 




00 Nd? 


(3926.123) \% 




^ 


3935.463 


Co,Fe 


I 


3926.165 J 




13 


3935.588 


,Mn? 


00 


3926.320 






000 N 


3935.787 




oNd? 


3926.465 


Ti 







3935.965 


Fc 


2 


3926.597 


Mn 




2N 


3936.121 


Co 


2 


3926.779 


Cr 







3936.507 




00 


3926.917 






000 


3936.699 







3927.079 






oNd? 


3936.912 


Mn- 


oNd? 


3927.269 









3937.105 




ooNd? 


3927.392 






000 


3937.283 




ooNd? 


3927.484 






00 


3937.479 s 


Fc 


3 


3927.585 






I 


3937.580 




000 


3927.748 









3937.695 




00 


3927.937 






iNd? 


3937.972 


Mn- 


ooN 


3928.075 s 


Fc 




8 


3938.116 







3928.231 


Fe 




2 


3938.160 




I 


3928.3^7 


-Cr? 




2Nd? 


3938.326 




000 


3928.485 






2N 


3938.439 




2 


3928.636 






00 


3938.552 




4 


3928.783 


Cr 




3 


3938.772 







3928.904 






000 


3938.876 




00 


3929.124 






000 N 


3939.007 







3929.260 


Fc-Co 




2 


3939.112 




00 


3929.363 


Fe-La-Mn 




2 


3939.288 







3929.497 


Co? 




I 


3939.532 




ooN 


3929.663 






oN 


3939.659 




00 


3929.864 


Mn? 




oNd? 


3939.736 




00 


3930.022 


Ti 




2 


3940.026 




ooN 


3930.180 






oN 


3940.183 


Fe? 


2 


3930.291 






000 


3940.324 




00 


3930.450 


Fc 




8 


3940.499 




oNd? 



TABLE OF SOLAR SPECTRUM WA VE-LENGTHS . 133 







Itttentity 






Intensity 


Wave-length 


Sabstance 


Mid 

Character 


Wave-leaeth 


Sobttaaoe 


and 

Character 


3940.812 






l5 


3947.300 




00 


3941.025 S 


Fc,Co 


3947.522 




2 


3941.190 







3947.67s 


Fc 


4 


3941.323 




I 


3947.830 




I 


3941.424 


Fc? 


3 


3947.918 


Ti 


2 


3941.510 




000 


3948.117 




000 


3941.637 


Cr 


3 


3948.246 


Fc 


5 


3941.753 




00 


3948420 




I 


3941.878 


Co 


3 


3948.613 




ooN 


3941.997 




I 


3948.818 


Ti 


4 


3942.157 


Mn?. 


oN 


3948.925 


Fc 


4 


3942.296 


Ce? 


1 


3949.039 s 


Ca 


I 


3942.380 







3949.199 


La 


I 


3942.510 ^ 




(2 


3949.286 


Fc 


2 


<3942.558) U 




< 


3949.372 




I 


3942.586 J 


Fc 


u 


3949-544 




00 


3942.747 




000 


3949.753 







3942.886 


CcCc 





3949.959 







3942.984 


Mn 





3950.102 s 


Fc 


5 


3943.238 




2 


3950.278 







3943.322 




2 


3950.398 




000 


3943^89 


Fc 


3 


3950.497 s 


Y 


2 


3943.622 




ON 


3950.613 







3943.721 




I 


3950.938 




000 


3943.819 




00 


3951.219 


Cr 


I 


3943.960 




(i 


3951.311 


Fc 


5 


3944.058 







3951.449 




00 


3944.160 s 


Al 


\ "5 


3951.578 




00 


3944.319 




liN 


3951.765 


Y 


oNd? 


3944.492 







3951.917 


Cr 


oN 


3944.681 




ooN 


3951.978 




000 


3944.824 




I 


3952.103 


Mn- 


2 


3944.884 


Fc? 


2 


3952.235 




00 


3945.033 


Fc 


3 


3952.342 


Na? 





3945.128 


Co 


I 


3952.465 


Co 





3945.260 


Fc 


3 


3952.549 


Cr 





3945.358 




r 


3952.606 




000 


3945.473 


Co 


3N 


3952.683 




000 


3945.633 


Cr? 


00 


3952.754 


Fc 


4 


3945.827 




00 Nd? 


3952.850 


Fc- 


3 


3945.993 


Mn? 


I 


3952.894 







3946.101 


Cr? 


00 


3953-043 


Mn 


3 


3946.188 


Fc 





3953.120 


Co 


3 


3946.340 







3953.222 




I 


3946.599 




00 


3953.303 


Fc-Cr 


3 


3946.693 







3953.400 







3946.800 


Co 





3953.550 




00 


3946.953 




oooN 


3953.641 




1 


3947.142 


Fc 


3 


3953.804 




000 


3947.272 


Co 





3953.844 







134 



HENRY A. ROWLAND 



Wsve-length 



3954.002 S 

3954.104 
3954190 
3954.414 
3954.536 
3954.680 
3954.857 
3954.967 
3955.157 
3955.356 
3955.482 
3955.524 
3955.744 
3955.903 
3955.968 

3956.099 
3956.197 
3956.316 
3956.476 
3956.603 
3956.819 
3957.030 

3957.177 S 

3957.423 
3957.621 
3957.767 
3957.939 
3958.073 
3958.231 
3958.35s 
3958.474 
3958.554 
3958.647 
3958.776 
3958.877 
3959.006 
39S9.I02 
3959.335 
3959.435 
3959.588 
3959.678 
3959.862 
3959.972 
3960.288 

3960.422 S 

3960.547 
3960.783 
3960.902 
3961.051 
3961.148 



Fc- 



Ni,Mn 
Fc? 



-Fe 



Fe 



Co-Ti 
Fc 
Fe 

Fc-Ca 



Fe,Co 

Co 
Ti, Zr 

Fc? 

Fe 



Fe 

Cr 

Co 



IntcBsity 

Mid 

Chancier 



3 

000 

ON 

000 

00 

2 

I 

000 

oooN 

I 

5 

00 

00 

o 

000 

3 

00 

00 

4 

4 

6 

00 

7<i? 

00 

ooN 

I 

00 

2 

00 

5 

000 

o 

000 

00 

I 

000 

00 

00 



00 

00 

000 

I 

00 

4 

000 

00 

00 

00 

o 



Wave-length 



3961.281 

3961.422 

396X.538 

3961.674 S 

3961.947 

3962.056 

3962.222 

3962.320 

3962.489 

3962.534 

3962.785 

3962.861 

3962.995 

3963.108 

3963.252 

3963.369 

3963.487 

3963.576 

3963.701 

.3963.831 

3963.941 

3964.059 

3964.173 

3964.326 

3964.416 

3964.541 

3964.663 

3964.897 

3965.148 

3965.366 

3965.487 

3965.614 

3965.655 

3965.750 

3965.868 

3965.980 

3966.069 

3966.212 

3966.494 

3966.647 

3966.778 

3066.966 

3967.194 

3967.570 

3967.777 

3968.000 

3968.114 

3968.350 

3968.625 s H 

3968.854 



Sabttanoe 



Fe 

Al 

Fe? 

Fc? 

Ti 
Fc 

Cr 

Ti 

Fe 
Fe? 
Co,. 
-Co 

Fe,. 

Fc 

Ni.Fe 
Fe 
Fe 

Fe 

Fe 

Ca 
Fe? 



Character 



3 

ON 

iN 

20 



I 

I 

2 

2 



00 



3 
00 

3 

I 

00 

I 

00 

3 

000 

00 

o 

I 

2 

00 

3 

oNd? 

oN 

o 

000 

000 

a 



I 

I 

2 

3 
oN 

2 

3 

I 

oNd? 

2 

oN 

oN 

I 

6N 

700 

iN 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 1 35 



W«ve*k^^ 



3968 J86 
3969.073 
3969.287 
3969.413 
3969.544 
3969.642 
3969.784 
3969.899 
3970.065 

3970.177 
3970.305 
3970.419 
3970.540 
3970.631 
3970.711 
3970.803 

3970.979 
3971.147 
3971.265 
3971.402 

3971.475 s 

3971.608 

3971.751 
3971.863 
3971.969 
3972.134 
3972.313 
3972.407 
3972.558 
3972.598 
3972.720 
3972.830 
3973.056 
3973.152 
3973.262 
3973.308 
3973.418 
3973.558 
3973.702 
3973.796 ) 

3973.864 5 

3974.004 

3974.057 

3974-164 

3974.310 

3974.408 

3974.536 

3974.637 

3974.774 

3974.904 



Cr, Co 
Fe 



Fc 
Cr 
Fc? 



Cr 
Fc 



Fc? 

Ni 

Co 

Cr? 
Fc 

Co 
Co 
Fc? 

Ni.Zr 
Fc 
Ca 

Fe? 



Fe 

Ni? 

Ni 

Co-Fc 



Intensity 
and 



6N 

oN 
oN 
10 
ON 

ON 

'2 
oN 
5N 

LoN 
I 
2 
I 

000 

o 

iN 
000 
o 

5 

000 

000 

00 

o 

oNd? 

2 

I 

I 



I 

00 

I 

00 

I 

I 

o 

ooN 

3 

w 

o 
I 

000 

I 

00 

3 

2 
2 
6d? 



Wave-length 



3975.197 
3975350 
3975.506 
3975.662 
3975.831 
3975.985 
3976.097 
3976.230 

3976.325 
3976.416 
3976.532 
3976.692 
3976.770 
3976.839 
3977.009 
3977.126 
3977.223 
3977.337 
3977.477 
3977.716 

3977.891 s 

3978.036 
3978.161 
3978.306 
3978.485 
3978.604 
3978.715 
3978.809 
3978.916 
3979.003 
3979.152 
3979.256 
3979.344 
3979.466 
3979.664 
3979.783 
3979.936 
3980.043 
3980.153 
3980.289 
3980.441 
3980.664 
3980.779 
3980.963 
3981.024 
3981.122 
3981.248 
3981.376 
3981.467 
3981.592 



Fc 
Co- 



Fc-Mn 



Fe 
Fe 
Fe 
Cr 
Fe 

Mn 
Co 



Fc 

Fc 

Co,Cr 

Co? 



-, Co 

Fc 

Fe,Cr 



Fe 



Ce 
Fe 
Cr 



Intensity 
and 



OONd? 

2 
I 

ooNd? 

ooNd? 

2 

000 

2 

000 

00 

2 

2 

3 

3 

2 

o 

00 



000 

ooN 

6 

000 

00 



I 

2 

00 

3 

000 

o 

000 

00 

00 

00 

4 

3 



00 

I 

00 

00 

00 

I 

000 

00 

I 

2 

I 

000 

000 



' The Hydrogen line must be wide and difiFuse, and thus probably coincides with 
this solar line which is of that nature. 



136 



HENRY A. ROWLAND 







Intensity 






Intenaity 


Wa^.lencth 


SabManoe 


•nd 

Character 


Wave-length 


Substanoe 


and 

Character 


3981.662 




00 


3988.617^ 




/' 


3981.762 


Zr? 





3988.659 \ 


«La 


1*" 


3981.917 S 


Ti 


4 


3988.705 J 




^"^ ^ 


3982.142 


Fc? 


2 


3988.812 




ooN 


3982.308 




ooN 


3988.979 




ooNd? 


3982.470 




ooN 


3989.137 




3 


3982.630 


Ti-Mn 


2 


3989.232 




2 


3982.742 


Y 


3 


3989.404 




ooN 


3982.896 




poo 


3989.592 




ooN 


3983.053 


Mn 


00 


3989.753 




ooN 


3983.150 




2N 


3989.912 


Ti 


4 


3983.341 




2N 


3990.011 


Fc 


3 


3983.503 







3990.129 


Cr-Mn 


I 


3983.682 




I 


3990.248 







3983.811 




00 


3990.333 




00 


3983.959 







3990.445 


Co 





3984.059 1 


Cr 


{l 


3990.525 


Fe 


2 


(3984.091) J-s 






3990.712 


V 


oN.d? 


3984.113 J 


Fc 


14 


3990.905 




oooN 


3984.294 


Mn 


2 


399I.O9S 




ooNd? 


3984.479 


Cr 


2 


3991.333 


Cr.Zr 


3 „ 


3984.592 




00 


3991.459 




ooN 


3984.707 




000 


3991.580 


Fc 


1 


3984.806 


Cc-Zr 


2 


3991.690 


Co 





3984.982 




00 


3991.830 


Co-Cr 


2 


3985.084 




I 


3991.892 







3985.213 




000 


3991.981 




00 


3985.385 







3992.O4X 




00 


3985.463 


Mn 


I 


3992.164 




000 


3985.539 


Fc 


5 


3992.261 




00 


3985.745 




00 


3992.396 




2 


3985.773 




00 


3992.538 


Fc 


I 


3985.939 




00 Nd? 


3992.631 




00 


3986.147 




ooN 


3992.790 







3986.321 


Fc 


3 


3992.971 


V-Cr 


3d? 


3986.439 







3993.121 




00 


3986.513 




000 


3993.246 


Fc 


2 


3986.717 




000 


3993.451 




ooN 


3986.903 s 




6 


3993.616 


Ba? 


ooN 


3986.979 


Mn 


I 


3993.758 







3987. 140 1 




I 


3993.878 







3987.244 ^s 


•Mn 


3 


3993.978 




00 


3987.332 J 


Co 


I 


3994.092 


Cr 


I 


3987.520 







3994.160 


Ni 


I 


3987.625 


Mn? 





3994.265 


Fc 


4 


3987.755 


Ti? 


2 


3994.416 




000 


3987.885 




000 


3994.474 




000 


3988.114 







3994.612 




00 


3988.259 




oooN 


3994.660 


Zr-Co 


2N 


3988.485 




ooN 


3994.828 


Nd? 


2 



' The Lanthanum line is a compound line made up of two or three lines covering 
about the width of these three lines of the spectrum. 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 137 



Wave-kafth 



3994.958 
3995.096 
3995.217 
3995.352 
3995.463 
3995.586 
3995.769 
3995.899 
3996.009 
3996.140 
3996.364 
3996.410 
3996.498 
3996^2 
3996.752 
3996.845 
3996.935 
3997.004 
3997.115 
3997.258 
3997.365 
3997.547 
3997.638 
3997.757 
3997.895 
3998x53 
3998.205 

3998.417 
3998.620 
3998.790 
3998.893 
3998.999 
3999.117 
3999.197 
3999.295 
3999.393 
3999.495 
3999.646 
3999.818 
3999.943 
4000.173 
4000.296 
4000.403 
4000.521 
4000.611 
4000.729 
4000.965 
4001. 119 
4001.256 
4001.315 



Ti 



Co 

Ba? 
La 

Fc 



Sc 



Fc 
Cr? 
Mn 
Fc 



Co 
Fc 



Ti 



Zr.Fc 
Cc 



Fc? 

Cr? 

Fc? 

Fc 

Fc 



Intensity 
•nd 

Chancier 



00 
00 
00 

2 

5 

00 
000 

I Nd? 
00 

3 



I 

00 

00 

00 

00 



000 

2 

I 

000 

4 
2 
000 

ooN 

4d? 

4 
000 

o 

4 

000 

000 

I 



000 



00 

000 

00 

000 

o 

000 

2 

o 

2 
00 

o 

000 


3 



WB««-leiwth 



4001.387 
4001.489 

4001.595 
4001.704 
4001.814 
4001.895 
4002.086 
4002.227 
4002.308 

4002.547 
4002.652 
4002.808 
4002.948 
4003.076 
4003.230 
4003.424 
4003.655 
4003.772 

4003.912 s 

4004.062 

4004.168 

4004.308 

4004.412 

4004.536 
4004.753 
4004.855^ 

4004.982 
4005.067 
4005.202 
4005.216 
4005.308 
4005.408 
4005.545 
4005.632 
4005.802 
4005.856 
4005.994 

4006.114 
4006.168 
4006.304 
4006.464 
4006.621 
4006.776 
4006.901 
4006.978 
4007.142 
4007.185 
4007.310 
4007.429 
4007.586 



^s 



Subitnnoe 



Mn 
Cr 
Fc 
Mn 
Mn 
Fc-Ti 



Cc-Fc-Ti 



Fc 



Fc 



Ti 

Nl 
Fc 

Fc 



Mn 
Fc 



Intensity 
•nd 

Chnracier 



000 
000 

I 

00 

3 

00 

ooNd? 



00 

00 

od 



00 

2 

000 

ooN 

ooN 

ooN 

3 

00 

o 

00 

000 

00 

00 

000 

I 

I 

I 

I 

oN 

7 
I 
I 
000 

3 
000 



00 

I 
2 
00 

3 

2 

I 

I 

00 

000 

3 
00 



138 



HENRY A. ROWLAND 



Wave-length 



4007.758 
4007.948 
4008.075 
4008.215 
4008.322 
4008.507 
4008.568 
4008.748 
4008.882 
4009.022 
4009.079 
4009.201 
4009.291 
4009.401 
4009.567 
4009.694 
4009.807 
4009.864 
4010.057 
4010.134 
4010.201 
4010.327 

4010.434 
4010.527 
4010.641 
4010.735 
4010.797 
4010.927 
4011.080 
40II.22I 

4011.451 
4011.561 
4011.693 
4011.865 
4012.046 
4012.174 
4012.305 
4012.399 
4012.541 
4012.631 
4012.757 
4012.858 
4012.941 
4013.107 
4013.221 

4013387 
4013.614 

4013.729 
4013.798 
4013.964 



Fc 

Co? 
Mn-Ti 



Ti 



Fc 

Ti 
Fc 

Ni 

Fc 



Fc 
Fc 



Fc 

Mn 

Fc 



Fc 

Nd,Zr 

Ti 

Cr 



Ti-Fc 
Fc 



00 Nd? 

00 

oN 

o 

o 

000 
000 

00 

2 

3 

000 

I 

000 
000 

I 
I 
3 

000 

00 

000 

I 

000 

00 

000 

3 

000 

2 
2 

00 
00 

3 

00 
2 

00 
00 

o 
I 

4 


ooNd? 

00 
00 
00 
00 

o 

000 

o 
3 
5 



Wave-leiwth 



OI4.II7 
014.274 

014.420 

014.537 
014.677 
014.827 

014.941 
015.086 

015.305 
015.416 
015.532 
015.628 
015.760 
io 1 5.880 
io 1 6.027 
iOi6.i54 
io 1 6.240 
iOi6.434 
iOi6.574 s 
io 1 6.695 
iOi6.835 

016.955 
OI7.II4 
017.244 
017.308 
017.458 

,017.620 
017.724 
017.925 
018.080 
018.234 
018.269 
018.420 

018.534 
018.646 
^018.840 
,018.996 
019.087 
,019.201 
io 1 9.287 

019.450 

io 19.580 

019.747 
019.987 

020.034 
020.171 
020.226 
020.341 
020.424 
020.547 



SobitaBoe 



Co 
Fc 
Fc 

Cc 

Ti? 



Fc 
Fc 

Co-Cr 



Fc 

Ni? 
Ni? 
Ti 

Mn 
Mn 
Fc 



NiCc 
Co 

Fc 
Mn 

So 



Intenaitj 



ij 



ooN 

ooN 

I 

00 

5d? 

00 

000 

oNd? 

o 

000 





3 

00 

00 

00 

000 



2 

00 

00 



00 

3 

4 

000 

2 

I 

O 

00 

3 



3 

00 

00 

000 

00 

00 

I 

00 

o 

000 

ooN 

000 

000 

o 

I 

I 

2 
I 



' Probably compound lines, of which there are many in the spectrmn of Manganese. 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 1 39 







Intemity 






Intensity 


WB««-length 


Subttanoe 


Mid 

Character 


Wave-lencth 


Sobttaaoe 


and 

Character 


402a639 


Fe 


1 


4027.539 







4020.799 




000 


4027.623 




00 


4020.927 




000 


4027.822 




I 


4021.057 


Co 


3 


4027.939 




00 


4021.2^8 




000 


4028.092 







4021.346 




000 


4028.272 




000 


4021.493 


Nd 





4028.497 


Ti- 


4 


4021.645 




000 


4028.638 




00 


4021.769 


Fe 


2 


4028.756 




00 


4021.893 


Ti 





4028.912 


Fe-Cc 


I 


4022.018 


Fc 


5 


4029.138 


Zr 


000 


4022.199 




ON 


4029.319 




000 


4022.371 


Fc 


I 


4029445 




000 


4022.401 


Cr 





4029.597 




ON 


4022.599 


Fe 





4029.796 8 


Fc-Zr 


5 


4022.685 


Cu? 


000 


4030.026 




ooN 


4022.775 




00 


4030.198 


Zr 


ooN 


4022.893 


Fe 


2 


4030.339 


Fe 


2 


4023.165 


Nd 





4030.497 


Sr 





4023.379 




ooN 


4030.646 


Fe-Ti 


5 


4023.533 


Co- 


3 


4030.796 







4023.705 


La 


000 


4030.878 ^ 


Mn 


U 


4023.834 ^ 


Sc 


2 


(4030.918) >8 




A 


4023.979 


Fc 


000 


4030.947 J 


Mn 


15 


4024.153 


Zr 


I 


4031.048 




2 


4024.247 


Fe 


2 


4031.270 







4024.373 




000 


4031.395 


Fe 


I 


4024.479 




000 


4031.492 







4024.593 




00 


4031.596 




00 


4024.726 


Ti 


3 


4031.712 




00 


4024.881 


Fe 


4 


4031.865 


Fe-La 


2 


4025.043 




oooN 


4031.942 


Mn 


2 


4025.158 


Cr 





4032.117 


Fe 


2 


4025.286 


Ti- 


3 


4032.265 




00 


4025458 




000 


4032.418 




00 


4025.579 


er 


I 


4032.610 


Fe 


2 


4025.733 




000 


4032.789 


Fe 


4 


4025.826 




000 


4032.985 




000 


4025.97a 


Co-La 


2 


403^112 




00 


4026.079 




00 


4033.224 8 


FeMn 


«7d? 


4026.220 




00 


4033.337 




I 


4026.318 


Cr 





4033.425 




00 


4026.461 




00 


4033.578 




oood ? 


4026.583 


Mn 


«2N 


4033.732 


Mn 


, 1 


4026.691 


Ti 


I 


4033.814 


Mn 


1 


4026.919 




00 


4033.946 




00 


4027.067 




00 


4034.052 




00 


4027.189 


Co 


I 


4034.120 




00 


4027.253 


Cr 





4034.245 




00 


4027.399 


Zr 


00 


4034.380 




iN 



' Probably compound Une8, of which there are many in the splctmm of Manganese. 



140 



HENRY A. ROWLAND 







InieodtT 






laieuity 


Wave-kaeA 


SnbMaBoe 


Character 


Wm-karh 


SdfaMDoa 


ud 


4034.533 




I 


4042.064 




00 


4034.644 s 


Mn-Fe 


-6d? 


4042.137 




000 


4034.880 




00 


4042.297 




000 


4035.018 




ooNd? 


4042.397 




00 


4035.265 




ooN 


4042.511 




000 


4035.399 







4042.591 




000 


4035.575 




00 


4042.743 


Cr.Nd 





4035.695 







4042.909 




00 


4035.752 


Co 


2 


4043.054 


La 





4035.883 s 


Mn 


Md? 


4043.145 




000 


4035.976 


Ti,Zr 


00 


4043.500 




ON 


4036.131 







4043.757 




00 


4036.251 




000 


4043.839 







4036.293 




00 


4043.956 


Cr 





4036.522 







4044.056 


Fe 


3 


4036.717 




00 


4044.141 




2 


4036.813 




000 


4044.2948 


K 





4036.923 




I 


4044.423 




000 


4037.076 




ooNd? 


4044.531 




00 


4037.268 




2 


4044.644 




[ 


4037.449 


Cr 


00 


4044.766 


Fe 


3 


4037.585 




000 


4044.992 




000 


4037.691 




00 


4045.108 




000 


4037.837 




I 


4045.266 


Mn 


f«i^ 


4038.094 




ooNd? 


4045.371 


Mn 


iiN 


4038.272 







4045.538 


Co 


5 


4038.425 







4045.662 







4038.627 




ooNd? 


4045.748 




2 


4038.771 


Fe,Mn 


I 


4045.864 







4038.944 




2 


4045.975 • 


Fe 


\ 30 


4039.094 




00 


4046.117 




iN 


4039.244 


Cr 


I 


4046.230 




2 


4039.444 




oN 


4046.490 




ooN 


4039.580 




000 


4046.612 


Cr 





4039.727 







4046.764 




000 


4039.891 




ooN 


4046.917 




oN 


4040.013 


Y 


00 


4047.171 




ooN 


4040.093 


Fc 


I 


4047.338 


K? 


ooNd? 


4040.243 


Fc? 


2 


4047.461 


Fe 


2 


4040.411 




00 


4047.556 




00 


4040.464 







4047.823 


Y 


ON 


4040.657 




ooN 


4047.958 




oN 


4040.792 


Fe 


3 


4048.224 




I N 


4040.937 


Cc, Nd. Co 


Id? 


4048.384 




00 


4041.099 




000 


4048.549 




ooN 


4041.221 




00 


4048.704 




00 


4041.431 


Fc 


3 


4048.818 1 


Zr 


[I 


4041.525 


Mn 


5 


(4048.883) Vi 




\ 


4041.803 


Zr. 




4048.910 J 


Mn-Cr 


15 


4041.957 




000 


4049.148 


Mn 






' Probably compound lines, of which there are many in the spectrum of Manganese. 



TABLE OF SOLAR SPECTRUM WAVELENGTHS 



141 



w« 



4049.303 
4049.363 
4049^82 
4049.590 
4049.716 
4049.882 
4050.019 
4050.176 
4050.254 
4050.482 

4050.643 
4050.716 
4050.830 
4050.963 
4051.095 
4051.204 
4051.336 
4051.491 
4051.742 
4051.888 
4052.070 
4052.176 
4052.316 

4052.454 
4052.603 
4052.650 
4052.812 
4052.871 
4052.992 
4053.091 
4053-263 
4053.424 
4053.582 
4053.639 
4053.839 
4053.981 
4054.085 
4054.225 
4054.335 
4054.457 
4054.591 
4054.714 
4054.863 
4054.962 
4055.023 
4055.189 
4055.365 
4055.533 
4055.701 s 

4055.855 



fSobitsnoc 



Fe 



La 
Zr 

Zr 
Fe 



V 

CrV 

Mn 
Fc 



Fe 
Mn 
Fe 
Fe 



Co 
Fe 

Fe-Ti 

Fe 

Zr 

Sc 



Fe 

Ti-Fc 

Mn 

Mn 





000 

2 

00 

iNd? 

I Nd? 

00 

00 

000 



00 

00 

2 

000 

00 

ooN 

00 Nd? 

oNd? 

000 

00 Nd? 

3 



00 

2 

2 

3 
I 

I 

000 
o 
00 

2 
o 

00 
00 

3 

000 

o 

I 

000 


ooN 
o 

2 

3 

3 

00 Nd? 

00 Nd? 

6 

00 



Wave-length 



4056.007 

4056.135 
4056.221 

4056.345 
4056.49s 
4056.601 
4056.708 
4056.955 
4057.055 
4057.225 
4057.368 
4057499 
4057.668 

4057.817 
4057.881 
4057.961 
4058.041 
4058.115 
4058.372 
4058.539 
4058.615 
4058.748 
4058.915 
4059.081 
4059.239 
4059.373 
4059.535 
4059.653 
4059.768 
4059.872 
4060.117 
4060.248 
4060.415 
4060.643 

4060.777 
4060.919 
4061.081 
4061.244 
4061.450 

4061.595 

4061.881 

4062.105 

4062.197 

4062.385 

4062.469 

4062.599 s 

4062.789 

4062.895 

4063.105 

4063.260 



Fc 
Cr 

Fc 



Co 
Fc 



Pb 

Mn 

Co-Fe 



Cc 

Fc, Cr 
Mn 



Mn 

Fe 

Ti 

Nd- 
Mn 



Fc 
Cu? 



Intensity 

and 
Character 



000 

O 
O 

I 

00 

oooN 

ooN 

00 

iN 

3 

7 



000 

000 





4 

000 
00 


3 
3 

00 
o 

«i N d? 
o 

000 
2 

00 
00 
I 
o 

00 

o 

00 

3 

oooN 
oN 
M2Nd 
2 

00 
00 
000 

5 


00 
00 
00 



' Probably compound lines, of which there arc many in the spectrum of Manganese 



142 



HENRY A. ROWLAND 



Wav«-Iength 



4063.436 

4063.573 

4063.759 » 

4063.938 

4064.075 

4064.201 

4064.362 

4064.515 
4064.607 
4064.728 

4064.913 
4065.239 
4065.388 
4065.537 
4065.737 
4065.861 
4065.960 
4066.155 
4066.270 
4066.375 
4066.524 
4066.742 
4066.870 
4066.975 
4067.139 
4067.429 
4067.558 
4067.642 

4067.753 
4067.915 
4068.005 
4068.137 
4068.268 
4068.488 
4068.694 
4068.801 

4068.999 
4069.115 
4069.221 
4069.306 
4069.423 
4069.588 
4069.761 
4069.890 
4070.055 

4070.195 
4070.431 
4070.589 
4070.777 
4070.930 



Fe 
Mn 
Fe 



Ti 
Ni 
Fe 

Lm 
Mn-Ti 

Fe 

Cr 



Co 
Fe 



Fe 
Fe 
La 



Fe-Mn 
Co 

Nd 

Mn 

Fe 



CharKter 



4 

20 

iNd? 

oN 

I 

I 

I 

2 

00 

000 

2d? 
o 

3 
oN 



000 

00 

I 
I 

2 
2 

o 



5 
3 

000 


o 
o 

00 

6 

000 
000 

o 

000 

00 

000 

2 
00 

o 

00 

I 

000 
000 

o 

3 

ooN 
ooN 
4 



inr avc * Wnfftll 



4071.137 

4071.252 
4071.501 
4071.680 
4071.789 

4071.908 s 

4072.059 

4072.121 

4072.295 

4072.508 

4072-655 

4072.853 

4073.052 

4073.287 

4073.493 

4073.637 

4073.780 

4073.921 s 
4074.203 
4074.488 
4074.681 

4074.835 
4074.947 
4075.055 
4075.257 
4075.468 
4075.661 
4075.857 
4075.995 

4076.101 
4076.201 
4076.283 
4076.375 

4076.516 
4076.644 
4076.792 

4076.959 
4077.033 

4077.221 

4077.348 
4077.498 
4077.630 

4077.731 

4077.885 s 

4077.985 

4078.126 

4078.318 
4078.515 
4078.631 
4078.801 



Character 



Zr 
Fe 

Fe 



Fe 
Zr 
Ce 



Ce 
Fe 



Fe 
Ni, Zr, Cr 

-Nd 
Nd 

Ce 
Ce 
Fe 
Cr 

Fe-Ce 

Fe-Zr 

Fe 
Fe 

Cr 

La.Y 



Sr 



Fe 

Ti 



{ 



00 

o 

000 

I 

o 

15 

oN 

000 

ooN 

oN 

2 

00 

ooNd? 

o 

000 

o 

000 

4 
00 

oN 
00 

2 

3 



2N 

oNd? 

000 

o 

00 

3 

00 

00 

I 
00 

2 

4 

2 

I 

oN 

00 

iNd? 

o 

oN 

8 

oN 

o 

ooNd? 

4 

3 

000 Nd? 



' Probably compound lines. 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 1 43 







iBtottity 






Intensitr 


WM«.Icagtli 


SolMlaiioe 


■Bd 

Chamcter 


Wave-length 


SobMiiioe 


■Bd 

Chancier 


4078.977 


Fc 





4086.861 


La 


I 


4079.161 




00 Nd? 


4086.988 




000 


4079.335 


Fc 


2 


4087.115 




000 


4079.393 


Mn 


3 


4087.252 


Fe 


3 


4079.508 




000 


4087.425 




00 


4079.570 


Mn 


3 


4087.482 




000 


4079.707 




000 


4087.641 




000 


4079.863 




00 


4087.755 




00 


4079.996 


Fe 


3 


4087.855 




000 


4080.209 




00 


4087.949 




00 


^^•3$! 


Fc,Nd 


3 


4088.200 




00 


4080.588 




00 


4088.333 




00 


4080.751 




00 


4088.448 




00 


4080.921 




00 


4088.596 




000 


4081.033 


Fe 


2 


4088.7131 


Fe 


3 


4081.190 




00 


4088.877 




00 


4081.385 


Zr,Ce 





4089.000 




00 


4081.415 


Fe 


I 


4089.199 




ooN 


4081.580 




ooN 


4089.374 


Fe 


3 


4081.736 




ooN 


4089.568 




oooN 


4081.887 




ooN 


4089.748 




oooN 


4082.060 




000 Nd? 


4089.935 




ooN 


4082.264 


Fe 


2 


4090.113 


Mn 


oNd? 


4082428 




oNd? 


4090.228 


Fe 


2 


4082.589 


Sc-Fc-Ti 


3 


4090.338 




000 


4082.749 


Co 





4090.474 


Cr 


ON 


4082.928 




ooNd? 


4090.671 


Zr 





4083.095 


Mn 


4 


4090.728 


V 


I 


4083.243 




000 


4090.921 


Zr 


00 Nd? 


4083.376 


Mii.Ce 





4091.109 




3 


4083.515 




000 


4091.235 




000 


4083.783 >* 


Fc 


1: 


4091.588 




oooN 


M11.Y 


4091.711 


Fe 


3 


4083.917 s 


Fe 


1 


4091.828 






4084.148 




oNd? 


4092.149 




00 


4084.476 




oN 


4092.248 




00 


4084.647 


Fe 


5 


4092.431 


Fc 


2 


4084.761 




000 


4092.547 


Co, Mn 


3 


4084.944 




oooNd? 


4092^5 


Fe 


I 


4085.161 


Fe 


4 


4092.821 


«V.Ca? 


3(i? 


4085.303 




00 


4092.975 




000 


4085.408 




1 


4093*041 


Co 


000 


4085.467 


Fe 


4 


4093.185 


Co 


00 


4085.59s 




000 


4093-435 




00 


4085.727 




00 


4093.801 




00 


4085.881 




00 


4094.141 




00 


4086.008 




ooN 


4094.220 




000 


4086.133 




I 


4094-455 




00 


4086.283 







4094.573 




2N 


4086469 


Co- 


3d? 


4094.761 




000 



144 



HENRY A. ROWLAND 







Intensitr 






iBmsity 


Wave-length 


ouimaiioc 


Chancier 


Wa«e-lei«th 


SubMlKC 


nd 
Chancier 


4094.849 




ON 


4102.541 


Y 





40950^4 


'Ca? 


4 


4102.774 




ooNd? 


4095.251 




000 


4102.914 




ooN 


4095.423 


Mn 





4103.097 s 


Si,Mn 


5 


4095-5" 







4103^70 







4095.633 


V 





4103.621 




00 


4095.795 




000 


4103.774 




oN 


4095.901 




000 


4103.967 




ooN 


4095.968 




000 


4I04.I4I 




000 


4096.129 


Fe 


3 


4104.288 


Fe 


5 


4096.262 


Fe 


2 


4104.459 




00 


4096.367 




I 


4104.623 


Co,V 





4096.481 




000 


4104.807 




000 


4096.675 




000 


4104.909 


Co.V 


00 


4096.795 




00 


4105.099 




I 


4096.848 


Zr? 





4105.216 




00 


4096.976 




000 


4105.318 


V 


2 


4097.101 




00 


4105.514 


Mn 


ooN 


4097.168 




00 


4105.807 




000 


4097.241 


Fe 


3 


4105.879 




000 


4097.389 




00 


4105.981 




000 


4097.612 




000 


4106.294 




000 


4097.733 




000 


4106.420 


Fe 


2 


4097.808 







4106.583 


Fe 


2 


4097.948 




00 


4106.738 




00 


4098.115 







4106.887 




oN 


4098.335 


Fe 


5 


4107.099 







4098.587 




000 


4107.449 




00 


4098.689 


'Ca? 


4 


4107.649 s 


Cc-FeZr 


5 


4098.746 




2 


4107.815 




000 


4098.947 




000 


4107.934 




00 


4099.054 




000 


4108.041 




00 


4099.207 







4I08.I8I 




00 


4099.327 


Ti 


00 


4108.289 




I 


4099.554 




000 


4108.454 




000 


4099.727 


La 


000 


4108.547 


Zr 


00 


4099.941 


V 


2 


4108.687 




2 


4100.147 







4109.062 







4100.315 


Fe 


2 


4109.215 


Fe 


3 


4100.501 







4109.374 




000 


4100.661 




000 


4109.609 


Nd? 


I 


4100.731 




000 


4109.734 


Cr 





4100.901 


Fe 


4 


4109.905 


•V 


2 


4101.067 







4109.953 


Fe 


3 


4101.244 







4IIO.II4 




000 


4IOI.42I 


Fe 


2 


41 10.194 




00 


4101.637 




fo 
3 


4110.454 




00 


4101.840 




4110.547 




00 


4102.000 Hd 


H, In 


^4oN 

Lo 


41 10.691 


Co 


4 


4102.321 


V 


4110.854 




000 



'These lines coincide with the heads of bands dde to Calcium. The bands proba- 
bly become lines owing to the weak dilation of Calcium vapor in the Sun. 



TABLE FOR SOLAR SPECTRUM WAVELENGTHS MS 



W«ve.|«ag,h 



4III.021 
4III.154 
41 1 1.35s 
41 1 1.509 
41 1 1.742 
41 1 1.827 
41 1 1.940 
41 12.139 
4113.234 
4112.327 
4112.478 
4112.603 
4112.721 
4112.869 
4112.959 
4113-067 

4113-117 
4113-247 

4113-381 

4113-679 
4113.839 
4114-021 
41 14-271 
41 14.461 

4114-606 S 

41 14.771 

4114.934 

4115-094 

41 15.189 

4115.330 

4115-533 

4II5.71S 

4115-834 

4115-961 

4116.047 

41 16.138 

4116.276 

4116.361 

41 16.474 

4116.634 

4116.707 

4116.859 

4116.974 

4II7.II3 

4117.317 

4117.414 

4117.587 

4117.741 

4117.894 

4M8.OO8 



Subttaaoe 



IniciMity 

and 
Character 



Mn 


I 


Mn? 


I 




000 


Cc? 


I 




000 




00 


'V 


4 




000 




00 




000 


Fe 


2 




000 




ooNd? 


Ti 


I 




000 




I 


Fe 


3 




00 


Mn 


I 




oooNd 




ooNd 


Mn 


ooNd? 




ood 


Mn 


000 


Fe 


4 




ooN 




ooN 




2 




000 


«V 


3 




ooN 




000 




000 




000 




00 









000 




000 




00 


«V 


I 


«V 





Nd? 


I 




00 









000 




00 




000 




00 




000 

2 



Wave- 



4II8.I48 
4118.307 
4118.347 
41 18.581 
4118.708 

4118.934 
4119.050 
4119.207 
4119.411 
4119.550 
4119.679 
4119.823 
4119.950 
4120.075 
4120.202 
4120.368 
4120.625 

4120.775 
4120.926 

4120.995 
4121.308 

4121.477 s 

4121.648 

4121.805 

4121.963 s 

4122.142 

4122.306 

4122.396 

4122.51 1 
4122.673 

4I22Jil9 
4122.936 
4123.027 
4123.184 
4123.384 
4123.430 
4123.539 
4123.664 

4123.713 
4123.907 
4124.030 
4124.097 
4124.260 
4124.359 
4124.510 
4124-645 
4124.782 
4124.938 
4125 067 
4125.285 



Fe 
Co 
Fe 



Fe 



Fe 



Cr-Co 
Zr 

Fe, Cr 

Ti,Cr 

Fe 



V 

Mn 

ti 

Fe 

Ce? 



Character 



00 


000 
000 

5 

4 

2 

oooN 

ooN 

I 



o 

o 

o 

000 

4 
000 

o 

000 
00 

00 

6d? 
ooN 
ooN 

3 

00 

I 

000 

ooN 

3 

I 

000 

000 

ooN 

12 

00 

o 

I 

000 

5 

00 
00 
000 

000 

000 

o 

000 

o 

o 

000 



-Th€»e mre some of the stronger Vanadium lines; and they are probably 
of prodiiciii^ the solar lines given in the Ubie. especially 41 ".940. 
•One of Uic strongest Lanthanum lines 



capable 



PHOTOGRAPHIC OBSERVATIONS OF ECLIPSES OF 
JUPITER'S SATELLITES.* 

By WiLLARD P. Gerrish. 

Since the year 1878 systematic observations have been made 
at the Harvard College Observatory of the eclipses of Jupiter's 
satellites. The variation of the apparent magnitude of the satel- 
lite, as it entered or emerged from the shadow, was determined 
by visual observations taken in rapid succession by means of a 
polarizing photometer attached to the 1 5-inch refractor. 

Since the recent successful applications of photography to 
photometric research, it was decided to make photographic obser- 
vations of all these eclipses visible at Cambridge, and the i i-inch 
Draper photographic telescope was chosen as being best adapted 
to the requirements of the work. The instrument is a visual 
refractor with a photographic correcting lens, the photographic 
focal length being about 144 inches. It is provided with an elec- 
trically controlled driving clock. It was found upon experiment 
thatan exposure of ten seconds with this instrument was sufficient 
to give satisfactory images of the satellites, and a plan of obser- 
vation was arranged upon that basis. The observation of a dis- 
appearance was conducted by exposing a plate upon Jupiter about 
eight minutes before the computed time of the eclipse, the tele- 
scope being kept directed by its driving clock. The slow 
motion in declination was then moved by hand at intervals of ten 
seconds, the time being taken from a chronometer. The motion 
was of sufficient rapidity to insure distinct, detached images of 
the satellites without the use of an exposing shutter, and was 
gauged to produce a displacement of the image on the plate of 
about 0.8 of a millimeter. This amount was doubled on the sixtieth 
second of each minute, thus dividing the chain of images into 
groups of six, each group representing one minute of time. The 

' Communicated by Edward C. Pickering, Director of Harvard College Observ- 
atory. 

146 



PLATE VIII 




Fig. I 



Fig. 2 



ECLIPSES OF JUPITER'S SA TELLITES 1 47 

exposures were continued for ten minutes in the case of the first 
and second satellites, the time being somewhat extended in the 
case of the third and fourth. Reappearances were observed in 
the same manner. By this method a continuous series of images 
was obtained, having a uniformity of exposure which admitted of 
direct photometric comparison, besides giving a faithful record of 
the beginning or ending of the period during which the satellite 
was too faint to be photographed. After completing an observa- 
tion, short independent series of exposures were made upon the 
same plate, with the aperture of the telescope reduced by known 
amounts. In this way a valuable check was obtained upon the 
scale of magnitudes used. 

The success of the photographic method just described sug- 
gested the use of an attachment which should be entirely auto- 
matic in its action, and the apparatus shown in the accompanying 
photograph (Plate VIII) was designed by the writer. A drum, 
four inches in diameter, is attached to the tangent screw of the 
slow motion in declination. Upon the drum is wound a cord, 
leading along the side of the telescope tube, through a pulley at 
the junction of the polar and declination axes, and thence to a 
weight of sufficient size to readily turn the drum. As the 
pulley is at the center of oscillation of the telescope, the 
telescope tube can be moved without affecting or being 
affected by the tension upon the cord. As the drum revolves 
under the influence of the weight, steel pins inserted in its 
inner edge engage successively the three-armed locking lever 
shown in the photograph. The upper end of the lever is 
held in position by a hook on the armature of a small electro- * 
magnet. The lever is similar in its action to an ordinary 
anchor clock-escapement. The drum turns in a direction corre- 
sponding to that of the hands of a watch, until stopped by a pin 
coming in contact with the left-hand arm of the locking lever. 
The end of this arm is slightly inclined, so that the lever, if free 
to swing, would be thrown aside, allowing the pin to pass. When 
an electric current is sent^ through the coils of the magnet the 
hook is lifted by the armature, freeing the lever, which is promptly 



148 WILLARD P. GERRISH 

thrown to the right by the action of the pin, thus unlocking the 
drum. The pin thus released passes along under the right-hand 
arm of the locking lever, returning the lever to its original posi- 
tion, where it is caught by the hook and is in readiness to receive 
the next pin. This device makes it possible to control a consid- 
erable weight with a small magnet and light current, though the 
pressure ordinarily exerted by the cord upon the drum is about 
one pound. The pins on the drum are six in number, and the 
spaces between them are equal, with the exception of that between 
the sixth and first, which is double that of the others, allowing 
the telescope to move twice the usual amount at the end of each 
minute, as already described. The electric signals operating the 
magnet are given at intervals of ten seconds by a special clock 
connection. In the small photograph is shown a portion of the 
photographic record of an eclipse which occurred on December 
10, 1892. The telescope in this case was operated by hand. The 
new apparatus is designed to give a similar record. The images 
in the photograph are enlarged to about three times the scale of 
the originals. 

Harvard College Observatory, 
December 14, 1894. 



THE ARC-SPECTRA OF THE ELEMENTS. H. 

By Henry A. Rowland and Robert R. Tatnall. 

GERMANIUM. 
(«f.-/. 9300 to 4600,) 



Wave * Imiftii 



CharKter 



Remarks 



iBtCIMtty 

in 
Sub 



2314.305 
2328.014 
2338-732 
2379.234 
2394.185 
2397.999 
2417.450 
2498.081 
2533.331 
2556.404 
2589.274 
2592.636 

2644.297 
2651.219 
2651.709 
2691.446 
2709.734 
2740.535 
2754.698 

2794.045 
2829.102 
3039.198 
3067.138 

3124.945 
3269.628 
3626.202 
4226.724 



4 

5 
2 
8 

2 

4 
20 
10 

18 
8 

18 

70 

12 
100 

80 

80 

90 

20 

90 

•5 

15 
100 r 

7 
25 « 

60 

4n 
7» 



Close to boron line. 



Also claimed by titanium. 
( Possibly a single reversed line. 



Coincides with ed^e of a broad solar line. 



Close to calcium line. 



5 
I? 



r indicates reversed, 

s indicates tkarp, 

n indicates haty or nedulaus, 

d indicates doubU. 

The germanium spectrum was obtained from argyrodite, 
which contains in addition to this element, silver, iron, mercury, 
zinc, and traces of indium, thallium and lead. The silver is very 
strong ; the others weak. 

149 



ISO H, A. ROWLAND AND R. R. TATNALL 

We give below a series of tables representing in detail the 
measurements made upon the argyrodite plates at our disposal, 
and from these the germanium lines given in the above list 
have been collected by a process of exclusion of the known 
substances. The tables give the results of nearly all meas- 
urements made on these plates, and contain, as will be seen» 
not only the lines of germanium and of other elements whose 
spectra are not well known, where these occur, but also the lines 
selected from the Table of Standard Wave-lengths for the 
purpose of establishing the scale and platting a correction-curve 
for each plate used. 

It was thought that the publication of these tables would be 
of use, inasmuch as they furnish direct comparisons between the 
wave-length measurements of lines due to different elements in 
the same region of the spectrum. Numerous iron lines have 
been measured on some of the plates, and these will give a 
ready means of comparison with the wave-length tables of 
Kayser and Runge, who use the iron-spectrum as their standard 
of reference. 

The second column in the tables contains the wave-lengths 
as read direct from the measuring engine. In spectra of the 
first order, these numbers are of course double the original 
readings, since the screw of the engine was made to correspond 
to the scale of second-order spectra. They are in many cases 
the means of several readings upon individual lines, or upon the 
lines and their ghosts, where these occur with sufficient distinct- 
ness to be used to advantage. The column headed '' Standard " 
contains the fractional part of the wave-lengths of standard 
lines, as taken from the Table of Standard Wave-lengths, and 
the correction-curve for the plate was platted from the weighted 
differences between these values and those obtained with the 
measuring-engine. Corrections for all the lines were then found 
from this curve, and appear in the column headed *' Correction." 
In the next column to this are given the corrected values of the 
wave-lengths, and where any line occurs on more than one plate, 
weights have been assigned, depending in general on the order 



ARC-SPECTRUM OF GERMANIUM 



151 



of the spectrum, the sharpness of the line, tts distance from the 
end of the plate, and to some extent on the character of the 
correction-curve. 



ARGYRODITE.— Plate I. 

FIRST ORDER SPECTRUM. 



Line 


W«ve-leagth 
(oMonected) 


Studanl 


Correctkm 


Wave-length 
(oonecied) 


Weight 


Ge 


2314.260 




- 


-.045 


2314.305 




Ge 


2327.976 






-.038 


2328.014 




Ge 


2338700 






•.032 


2338.732 




Al 


2367.146 


.144 




-.018 


2367.164 




Ge 


2379.220 






-.014 


2379.234 




In 


2389.570 






-.007 


2389.577 




Ge 


2394.180 






-.005 


2394.185 




Ge 


2397.996 






-.003 


2397.999 




In? 


2399*340 






-.003 


2399.343 




Ge 


2417.456 




—.006 


2417.450 




Si 


2435.252 


.247 


—.014 


2435.238 




C 


2478.680 


.661 


—.032 


2478.648 




B 


2496.896 


.867 


—.038 


2496858 




Ge 


2498.120 




—.039 


2498.081 




Si 


2519.348 


.297 


—.047 


2519.301 




Ge 


2533-384 




-.052 


2533.331 




Ge 


2556464 




—.060 


2556.404 




Tl 


2580.356 




-.067 


2580.289 




Ge 


2589.344 




— .070 


2589.274 




Ge 


2592.708 




—.072 


2592.636 




Fe 


2599.558 


.494 


—.072 


2599.486 




GeTi 


2644.380 




-.083 


2644.297 




iS 


2651.304 




—.085 


2651.219 




2651.794 




-.085 


2651.709 




Ge 


2691.536 




—.090 


2691.446 




Ge 


2709.824 




—.090 


2709.734 




Fe 


2719.218 


.119 


—.090 


2719.128 




Fe 


2721.080 


0.989 


—.090 


2720.990 




Fe 


2723.772 


.668 


—.090 


2723.682 




Ge 


2740.624 




-.080 


2740.535 




Ge 


2754.786 




—.088 


2754.698 




Tl 


2768.078 




-.087 


2767.991 




Ge 


2794.130 




—.085 


2794.045 




Mg 


2795.700 


.632 


-.085 


2795.615 




Mg 


2802.890 


.805 


-.084 


2802.806 




Ge 


2829.182 




—.080 


2829.102 




Mg 


2852.328 


.239 


—.077 


2852.251 




Si 


2881.760 


.695 


—.074 


2881.686 


I 


Tl 


2918 518 




—.068 


2918.450 




Fe 


2967.098 


.016 


-.057 


2967.041 


I 


Fe 


2973.384 


.358 


—.056 


2973.328 




Si 


2987.840 


.766 


— 053 


2987.787 




Ge 


3039.260 




— .040 


3039.220 


2 


Fe 


3047.774 


.720 


—.037 


3047.737 





COMPARISON OF PHOTOMETRIC MAGNITUDES OF 

THE STARS. 

By Edward C. Pickering. 

In the ninth volume of the publications of the Observatory at 
Potsdam the photometric magnitudes obtained at that institution 
are compared with the results of three other photometric cata- 
logues. These are the Uratiametria Oxaniensis^ the Harvard Photom-- 
etry {^Harvard Annals, Volume XIV) and the results contained in 
Volume XXIV of the Harvard Annals. From this it appears that 
the number of stars in each of these catalogues observed also at 
Potsdam is 691, 791, and 801 respectively. The number of cases 
in which the difference exceeds half a magnitude, after applying 
a correction for the systematic difference due to the color of the 
stars, is similarly 21,6, and 13 in the three cases. As this correc- 
tion is almost exactly the same for the Oxford and Harvard cata- 
logues, it has been found convenient in the following table to apply 
the correction for color to the Potsdam magnitudes. As no com- 
parisons are made in the Potsdam volume of the magnitudes of the 
stars observed both at Oxford and Harvard, it seems worth while 
to give them below, especially as later observations of many of 
these stars have also been obtained at Harvard in the re-observa- 
tion, now nearly completed, of all the stars contained in the Har- 
vard Photometry, 

In the following table the forty stars mentioned above, and 
enumerated in the Potsdam volume (page 500), are given on 
successive lines. The number from the Potsdam catalogue in the 
first column is followed by the magnitudes in the five catalogues 
mentioned above. They are arranged in chronological order, as 
the mean year of observation of the Harvard Annals, Volume XIV, 
is 1 88 1, of the Uranometria OxonUnsis, 1884, Harvard Annals, Vol- 
ume XXIV, 1885, Potsdam Observations, 1890, and the revision 
of the Harvard Photometry, 1893. '^^e means and residuals of 
these magnitudes are given in the subsequent columns when a 

154 



PHOTOMETRIC MAGNITUDES OF THE STARS 155 
Oxford, Potsdam, 





s88i 
H«iy'd 


S884 
0]?fd 


Har^d 


PttS^m 


J». 


Mctti 




Residualt 




XIV 




XXIV 




H.P. 










10 


3.04 


2.47 


2.77 


3.02 


a^5 


2.83 


+.21 


-.36 


-.06 


-J-.I9 


+.02 


318 


5.78 


5.72 




6.22 


6.14 


5.96 


-.18 


-.24 




--.26 


--.18 


332 


2.68 


2.44 


2.93 


2.93 


iXm 


2.76 


-.08 


-.32 


+.17 


--.17 


--.04 


638 


, , 


6.06 




7.22 


, , 




. . 










849 


5.66 


5.61 


, . 


6.12 


5.89 


5.82 


-.16 


—.21 




+.30 


+.07 


899 


4.40 


4.89 


4.38 


4.36 


4.13 


4.43 


-.03 


--.46 
--.46 


-.05 


-.07 


—•30 


984 


4.62 


5.19 


4.71 


4.63 


4.50 


4.73 


—.11 


— .02 


—.10 


--.23 


990 


5.62 


5.45 


6.02 


5.98 


5.91 


5.80 


-.18 


-.35 




-.22 




-.18 




-.11 


1754 


6.30 


6.19 


6.95 


6.71 


6.58 


6.55 


-.25 


-.36 




-.40 




-.16 




•03 


I8I3 


5*59 


5.65 


6.04 


6.18 


6,27 


5.95 


-.36 


-.30 




-.09 




-.23 




-.32 


I9I7 


6.61 


6.30 


6.74 


6.82 


6.81 


6.66 


-.05 


7.36 




-.08 


- 


-.16 




-15 


2171 


. . 


7.13 


, . 


6.56 


6.64 


6.78 




+.35 


. . 


—.22 


T'^ 


2610 




3.91 




4.73 


4.56 


4.40 




-.49 




+.33 


+.16 


261 1 




4.23 




5.16 
















2622 


5.24 


5.03 


5.57 


5.56 


5.44 


5.37 


-.13 


-.34 


--.20 
"IS 


- 


-.19 


+.07 


2756 


6.64 


6.31 


6.81 


6.88 


6.64 


6.66 


—.02 


-.35 


- 


-.22 


—.02 


2776 


5.69 


5.45 


5.91 


5.93 


5.74 


5.74 


-.05 


-.29 


--.17 


--.19 


.00 


3026 


4.13 


3.59 


3.80 


4.17 


3.99 


3.94 


+.19 


-.35 


-M 


- 


-.23 


+.05 


3327 


5.18 


4.99 


5-50 


5.50 


5.25 


5.28 


—.10 


-.29 


--.22 


- 


-.22 


-.03 


3343 


5.27 


5.05 


5.67 


5.78 


5.44 


5-44 


-17 


-.39 


--.23 


- 


-.34 


.00 


3414 


2.61 


2.33 


2.49 


2.94 


2.66 


2.61 


.00 


-.28 


— .12 


" 


h.33 


+.05 



Harvard XIV, Potsdam. 



1737 


6.75 


6.47 


6.30 


6.25 6.41 


6.44 


+.31 


+.03 


-.14 


-.19 


-.03 


I8I3 


5.59 


5.65 


6.04 


6.18 


6.27 


5.95 












2044 


5.71 


5.79 


. . 


6.25 


5.90 


5.91 


—.20 


—.12 




+.34 


—.01 


2326 


6.45 






5.58 


5.59 


5.87 


-.58 
".15 






-.29 


-.28 


2918 


6.01 


6.00 




5.50 


5.91 


5.86 


+.14 




-.36 


+.05 


3343 


5.27 


5.05 


5.67 


5.78 


5-44 


5-44 




•• 



















Harvard XXIV, 


Potsdam. 










122 


5.68 


5.60 


5.79 5.50 


5.59 


5-63 


+.05 


-03 


-h.i6 


-.«3 


-.04 


185 






6.21 


6.50 


6.30 


6.34 






, 


T-'3 


+.16 


-.04 


245 


5.55 


550 


6.15 


5.60 


5.52 


5.66 


-.11 


-.16 


+.49 


-.06 


-.14 


653 


6.49 






6.62 


6.22 


6.81 


6.54 


-.05 




, 


+.08 


-.32 


+.27 


1218 








7.71 


9.05 


. , 
















1698 








6.75 


7.02 




















1699 








7.70 


7.56 


, , 


, , 














, , 


1752 








6.93 


6.36 




















2395 








6.48 


5.98 


S.83 


6.10 










+.38 


—.12 


-.27 


2535 








8.20 


7.64 




















3094 








6.91 


7-73 




















3166 








8.07 


7.44 




, , 










, , 


, , 


, , 


3361 


3.59 


329 


3.78 


3.48 


3-4» 


3.51 


+.08 


— .2 


2 


+.27 


-.03 


-.09 



Average deviations . 



±.15 ±.29 ±.i8 db.2I H=.1I 



156 EDWARD C. PICKERING 

star appears in more than two catalogues. Stars 181 3 and 3343 
appear twice in the table, and the residuals for the second entry 
are therefore omitted. Stars 122, 185, 653, 3327, and 3361 occur 
in both Tables I and IV of Volume XXIV, and the mean of the 
magnitudes is accordingly entered in the table. 

From the average deviations given in the last line the order 
of excellence of the first three catalogues appears to be the same 
as that found at Potsdam, Harvard XIV first. Harvard XXIV 
second, and Oxford third. Moreover, the Potsdam observations 
are more accordant than the Oxford, and the recent Harvard 
work more accordant than any. The observations in Volume 
XIV were made with an instrument smaller and inferior to that 
used in the later Harvard work, but as the minimum number of 
nights' observations was three instead of two, the average devia- 
tions for Volumes XIV and XXIV do not differ greatly. It 
should be noticed that the above stars are in general the most 
discordant in the catalogues, as they include all the stars in which 
the results differ more than half a magnitude. The average devia- 
tions of the other stars would be mucA less. To treat the three 
catalogues, Oxford, Potsdam, and Harvard XIV symmetrically, 
stars differing in Oxford and Harvard XIV by half a magnitude 
should also be included. On examination it appeared that there 
were only two such stars, 984 and 3026, in the part of the sky 
here under discussion, and these are already included in the above 
list among the stars differing at Potsdam and Oxford. The Har- 
vard results agree more closely with both the Potsdam and the 
Oxford measures than the latter do with one another, the number 
of differences in the three cases exceeding half a magnitude being 
6, 2, and 21 respectively. In the average deviations given below 
it might be thought that the results were affected by errors per- 
sisting in all the measures made with the meridian photometer. 
This would not be likely to occur, as two different instruments 
were used, and an interval of several years occurred between the 
second and third series, during which the instrument was dis- 
mounted and sent to South America, and a long series of measures 
made with it there. There is no evidence of such a persistence 



PHOTOMETRIC MAGNITUDES OF THE STARS 1 57 

in the corresponding residuals. A recomputation has, however, 
been made, taking the mean of all the meridian photometer 
observations of each star, and giving equal weights to this mean, 
to the Potsdam, and to the Oxford results. The average devia- 
tions then become, for Potsdam^ .21, for Oxford^ .27 and for 
Harvard^ .11. The di£Eerences in the three Harvard catalogues 
separately become, for Volume XIV, . 1 7, Volume XXIV, .2 1 , and 
for the revision of the Harvard Photometry^ .12. The di£Eerences 
in the different catalogues cannot be mainly due to progressive 
changes in light of the stars, or we should expect the first and 
last catalogues to give the largest average deviations, also that the 
extreme values would occur most frequently in these catalogues. 
A count of these extreme values shows that the number for each 
catalogue is, for Harvard XIV, 6, for Oxford^ 20,*for Harvard 
XXIV, 10, for Potsdam^ 19, and for the revision of the Harvard 
Photometry, 5. The order is again the same as for the average 
deviations. 

The positions of 1698 and 1699 were interchanged in printing 
Volume XXIV. Their magnitudes as given in Volume XXIV 
have therefore been transposed in the above table, since the 
stars were entered correctly in the observing list, and correctly 
observed. In fact, on three of the four nights on which they 
were measured, the observer noted the position of the adjacent 
star. A large part of the errors of Volume XXIV, such as those 
relating to AZ>.+59** 76 {A.N. 136,85) are of this class, the stars 
being correctly observed but the places incorrectly printed. 
Slight errors in the Durchmusterung, especially in declination, were 
often noticed by the observer, but even an error of a minute of 
time, as in the case of ^.Z>.+45*' 921, probably seldom led to the 
observation of another star, especially when, like i9.Z>. +69° 455, 
it was near the pole. In this latter instance the error in the Dtirch- 
musterung of a minute of time only displaced the star 5 \ or about 
one-ninth of the diameter of the field. Moreover, on both nights 
of observation the observer recognized the error by noting that the 
star had the same right ascension as B. /?.+ 70° 497. Appar- 
ently some other star was observed by mistake for B. Z>.-f-5® 1934. 



158 EDIVARD C. PICKERING 

An attempt has been made throughout the Harvard work to 
eliminate systematic errors as much as possible, since if this is 
done the accidental errors can be reduced indefinitely by increas- 
ing the number of observations of each star. The accidental 
errors in either catalogue are of little importance, since they affect 
the final values by only a few hundredths of a magnitude, while 
the systematic differences amount in some cases to several tenths. 
At Harvard, when an observation was discordant, the star was 
re-observed on so many nights that the final value was generally 
only changed one or two tenths of a magnitude, whether the 
observation was retained or rejected. 

The Potsdam astronomers are in error in one respect, in sup* 
posing that the Harvard observations were made with undue 
haste. In (Quoting from Volume XXHI, p. 7, the statement of 
the speed attained, they have failed to quote the sentence follow- 
ing it. . "Care was always taken that the observer should not be 
hurried in his measures, and the work was accordingly so divided 
that he generally had to wait for the recorder." A star a minute 
is not an unusual speed for observations of stars in zones with 
meridian circles, where much more accurate measures are required, 
and is never regarded as an evidence of inaccuracy. Under favor- 
able circumstances stars can be observed at this rate with the 
meridian photometer, and the results will have an average devi- 
ation of almost exactly a tenth of a magnitude. Of course, 
several minutes are often spent upon a single star when adjacent 
stars render its identification uncertain. The great number of 
measurements obtained may be assigned to two causes. First, as 
in the case of other meridian instruments, very little time is lost 
in setting upon the star and identifying it with certainty, so that 
almost the entire time of the observer is spent in the actual pho- 
tometric comparisons. Secondly, by long practice observers and 
recorders are enabled to perform the mechanical operations with 
great rapidity. The average number of stars, not including stand- 
ards, observed each evening at Potsdam was eighteen, while at 
this Observatory, when the work is not interrupted by clouds, it 
now exceeds one hundred. To fairly compare the two methods 



PHOTOMETRIC MAGNITUDES OF THE STARS 1 59 

the error of the mean of five or six observations made here should 
therefore be compared with that of a single observation at Pots- 
dam. At Oxford ten stars were generally observed each night, 
four sets of measures being made of each star. The above dis- 
cussion with that given elsewhere shows that the work done with 
the meridian photometer compares favorably with that obtained 
by other methods, both as regards speed and accuracy. 
Harvard College Observatory. 



THE SPECTRUM OF S CEPHEL* 

By A. BiLOPOLSKY. 

Since August 3 I have been able to secure thirty-four spectro- 
grams of this star. As my observations were interrupted on Sep- 
tember 12, and may not be resumed for some time, I will give the 
results here, although they are to be regarded as only provisional. 

The star belongs to type Ua, like a Bootis. 

The prism was adjusted to minimum deviation for X 4410, and 
the displacements relatively to X 4410.5 and A 4405 were measured 
by Vogel's first method. The velocities, reduced to the Sun, are 
as follows (two spectrograms were obtained on each evening) : 



1894 


Pttlkows 
M. T. 


KOometen 


1894 


FuIlGOWA 

M. T. 


KikMneiera 


I 


Aug. 


3 


I|h 


— 32 


I0« 


Aug. 17 


II»» 


+ I 


2 


<i 


4 




— 23 


II 


" 24 


10 


— 39 


3 


M 


5 




— 4 


12 


" 25 


10 


-25 


4 


(« 


6 




+ 5 


13 


Sept. I 


9 


— 6 


5 


«« 


8 




— 30 


M 


" 3 


9 


-36 


6» 




9 


10 


— 22 


15 


" 5 


9 


— 22 


7 


«l 


12 




— 18 


16 


•* 6 


9 


— 9 


8 


»t 


14 




— 24 


17 


" 7 


9 


- 5 


9 


** 


16 




— 4 


18 


" II 


9 


— 13 



The epochs of the minima were, according to the Annuaire 
du Bureau des Longitudes^ July 31, i6\ Pulkowa mean time, Aug. 
6, I ^ Aug. 1 1, io\ Aug. 16, I9^ Aug. 22, 3\ Aug. 27, I2^ Sept. 
I, 21 ^ Sept. 7, 6\ 

With the velocities obtained I drew the curve of velocity (see 

R. Lehmann-Filh^s, A, N,^ 3242), and obtained for - | -^dt 
the value — i8*"".4, minus the velocity of the system. 
The following quantities are then obtained : 

Z,= + 2i.7, A = 21^, 

2^ =-58.7, B=20^, 

' Translated from A, M, 3257. German geographic miles in the original have 
been changed to kilometers. 
* One spectrogram. 

160 



from which 



THE SPECTRUM OF S CEPHEI 



= 90 



161 



aya^'-a. 



e = 0.46, 

7* = + I'^.o;, reckoned from the epoch 
of the light minimum, a sin i = 1,300,000*^, 

^ = 5«' 9*» (assumed). 
To what extent these elements correspond with the observa- 
tions, and how the velocity curve fulfils the necessary conditions, 
may be seen in the table which follows. The observed velocities 
were freed from the velocity of the system and arranged accord- 
ing to the argument of the latitude. 







Obierrcd 




4iM A +B A-B 


Na 


« 


Velodtj 


Owe 


- -37 =. OMW + 


II 


42^3 


-I9»» 


— 16 


— 16 


S 


48.1 


— 13 


-15 


— 14 


I 


63.3 


— 14 


— II 


— 10 


8 


71 .3 


— 6 


— 8 


— 7 


12 


78.7 


— 6 


— 5 


— 4 


6 


81 .3 


— 3 


— 4 


— 3 


15 


84.9 


- 3 


— 4 


— 2 


2 


92 .9 


— 4 


-I- I 


+ I 


18 


102 .5 


+ 5 


-f 4 


+ 4 


16 


"3 .5 


+ 10 


+ 8 


-h 8 


3 


"3 .3 


-f 14 


+ 11 


+ 11 


13 


126 .3 


+ 14 


+ 12 


+ 13 


9 


132 .7 


+ 5 


-f 13 


-f 14 


17 


154 .7 


+ 13 


+ 19 


+ 19 


4 


175 .7 


+ 24 


-I-21 


+ 21 


10 


201 .7 


4-20 


+ 20 


+ 19 


7 


265 .9 


+ I 


-f 2 


+ I 


14 


344 .1 


— 18 


— 19 


— 20 



The sign of the velocity is given according to the usual con- 
vention. 

It is to be noted that the epoch of minimum brightness occurs 
a day earlier than the time of perihelion passage, at least accord- 
ing to the ephemeris which was used. Whether these two ele- 
ments can be brought into accordance is a question for the future. 

PULKOWA, 

September, 1894. 



Minor Contributions and Notes. 



SPECTRO-BOLOGRAPHIC INVESTIGATIONS AT THE SMITH- 
SONIAN ASTROPHYSICAL OBSERVATORY. 

The bolometric researches described and illustrated by Professor 
S. P. Langley' at the Oxford meeting of the British Association are so 
important as to call for more than passing comment in The Astro- 
physical Journal. Some of the results obtained by the use of auto- 
matic methods of registering the indications of the bolometer are 
illustrated in Plate IX. 

The infra-red region of the spectrum has been the subject of many 
investigations since the early work of Sir John Herschel in 1840. His 
method of causing the solar spectrum to fall upon paper moistened 
with alcohol, and noting the rapidity with which the paper dried in 
various regions, was quite good enough to reveal to him the existence 
of cold bands in the spectrum (Fig. i, a). Later observers obtained 
results of great importance with the thermometer and thermopile, and 
with phosphorescent and photographic plates. Lamanski's curve, 
representing the distribution of energy in the infra-red solar spectrum, 
as determined in 187 1 by his obs'ervations with a thermopile, is illus- 
trated in Fig. I, b. The limitations of the method, due to the large 
surface and comparative insensitiveness of the thermopile, render the 
results crude as compared with those obtained later with the bolometer, 
but a distinct step in advance had been taken. The younger Becquerel's 
results, and particularly the fine map of the upper infra-red due to 
Lommel, seem to indicate that the phosphorescent plate method is of 
very great value. It can hardly be doubted that it is capable of still 
further improvement. 

But while the phosphorographs of Lommel and the photographs of 
Higgs are admirable in the upper infra-red, it is to the bolometer that 
we must look for a knowledge of the region lying beyond. The inves- 
tigations of Langley, Rubens, Snow and Paschen have made the 
bolometer a convenient and thoroughly practical instrument. Since 
the work of Langley at the Allegheny Observatory the improvement 

■"On Recent Researches in the Infra-red Spectrum," by S. P. Langley. 

162 



\ 



MINOR CONTRIBUTIONS AND NOTES 



163 



and increased sensitiveness of the apparatus has not been so evident 
in the bolometer itself, though it is true that this has undergone cer 
tain modifications for the better. But it is to improvement of the 
galvanometer that most attention has been directed. The immediate 
result has been most marked, as is indicated by the following table,' 
which shows the advances made in this direction at the Smithsonian 
Astrophysical Observatory in a single year. 





Description 


Old ooastant* 


New oonsttim 
ment) 


(Old) 


D'Arsonval 

Thomson 


0.00000010000 
0.00000000150 
0.00000000160 


.00000002000 


White (old) Allegheny pattern. . 
Queen (old) 


.00000000070 


Thomson 


.00000000040 


Elliott Bros, special design (new) 
Nalder Bros, special design (new) 
Nalder Bros, special design (new) 


Thomson 


Thomson (multiple)^ 
D'Arsonval 




.00000000002 




Not finished 









So high a degree of delicacy has thus been reached that at the pres- 
ent time the conditions of use at the Observatory are such as to render 
only about one-tenth of this increased delicacy available.^ 

Fig. I, c, shows the distribution of energy in the spectrum of a 60° 
prism of rock-salt, as determined by Professor Langley in his work at 
the Allegheny Observatory with a bolometer used in the ordinary way. 
Independent galvanometer deflections were measured with the bolom- 
eter strip set in various parts of the spectrum, and from the readings 
thus obtained the curve was plotted. The bolometer used was of such 
a degree of sensitiveness that it would indicate a change of temperature 
of \^^^^^ of a degree centigrade, and with it the error in determining 
the position of a line was within a minute of arc. The presence in the 
spectrum of the D line of sodium as a single line could barely be detected 
with this apparatus. The method employed made the work necessarily 
very slow, and so many independent observations of each line were called 
for that the positions of only twenty lines were recorded in two years. 
The recent work of Snow shows a marked improvement in this respect. 

In the method now employed at the Smithsonian Astrophysical 

' From the Annual Report of S. P, LangUy, Secretary of the Smithsonian Institu- 
tion^ i8g4, 

* Current which deflects image one millimeter at distance of I meter, when the 
time of a single vibration is 10 seconds. 

> Partially finished. 

4 The methods by which these results were obtained are described by Professor F. 
L. O. Wadsworth in the Philosophical Magauine for November and December, 1894. 



1 64 



MINOR CONTRIBUTIONS AND NOTES 



Observatory the time required to map a given region of the spectrum 
has been reduced to a minimum by a very ingenious contrivance. A 
''fixed-arm'' spectro-bolometer is employed in connection with a 
large Foucault siderostat. All of the prisms and objectives are of 




LAfiANSKfS CUf^C 




DI5THIBUTI0N «r CNCRfiY 

INTHS 

5P€CTFtUM 
or A60* 
PRISM or SALT 



f.6.1 

rock-salt. The azimuthal circle which carries the prism is in 
mechanical connection with a distant photographic plate, and 
both are moved by the same clockwork. The arrangement (in 
certain cases) is such that the plate is carried down vertically 
one centimeter while the circle and prism revolve through one 
minute of arc. It is obvious that the spot of light reflected from 



MINOR CONTRIBUTIONS AND NOTES 



165 



the galvanometer mirror on to the photographic plate will under these 
conditions trace a curve representing the variations in temperature of 
the bolometer strip. As the bolometer is placed at the end of a very 
long arm its angular width can be made quite small, and its position in 

the spectrum at any given time 
can be determined with a con- 
siderable degree of accuracy. 

Fig. 3, Plate IX, is a photo- 
graphic reproduction of three 
curves representing the entire 
region investigated. They were 
all made on the same day, and 
except for the variation in the 
Sun's altitude, and the interposi- 
tion of invisible clouds, they 
should show no differences other 
than those due to accidental dis- 
turbances of the apparatus. They 
are, in fact, quite concordant, and 
they clearly show that the auto- 
matic registration method has 
certain important and well-de- 
fined advantages. 

These three curves were pur- 
posely made with a rapid move- 
ment of the clockwork and brief 
swing of the galvanometer, and 
they do not represent all of the 
lines within reach of the bolom- 
eter. Under conditions of the 
highest delicacy both D, and D„ 
them, can be recorded, as Fig. 2 




Fig. 2 

nickel line between 



as well as the 
illustrates. 

The process employed in making the linear representations of the 
curves has not been described in its details, but it is easy to see how it 
could be carried out by the use of a suitable cylindrical lens com- 
bined with spherical lenses. For almost all scientific purposes the 
curves themselves would be of more interest and value than their 
linear representations. It may be doubted whether the composite 



l66 MINOR CONTRIBUTIONS AND N0TE6 

process of formins^ a single linear representation of a number of these 
curves would dispose of all accidental lines. It is probable that some 
of the accidental irregularities in the spectrum curve would appear as 
lines in the spectrum, and some false lines might be introduced from 
other sources. Thus the portion of the spectrum shown in Fig. 5» 
Plate IX, may contain a considerable number of lines not really 
present in the solar spectrum. In widening photographs of stellar 
spectra with a cylindrical lens many lines due to dust and irregu- 
larities in the film frequently appear in the enlargements, and it 
could hardly be expected that these could be altogether avoided in 
holographs. 

In spite of the great interest and value of the holographic method, 
it is hardly probable that it will altogether take the place of the ordi- 
nary process of uncovering the slit, and observing the galvanometer 
deflection. The latter method is much slower, but for detailed study 
of spectral lines it is in some respects superior, particularly as regards 
delicacy. For, unless the motion of the spectrum over the bolometer 
is very slow, the galvanometer needle will hardly have time to reach 
the full deflection. Again, the positions of lines can probably be 
determined rather more accurately by the ordinary method, on account 
of the inertia of the galvanometer needle and the unavoidable lost 
motion in the train of gearing connecting the azimuthal circle with 
the photographic plate. But for much work the ingenious automatic 
method has many advantages over the older and slower plan, and it 
will surely be a valuable acquisition to astrophysics. G. E. H. 



DEVICE FOR PUTTING WAVE-LENGTHS ON SPECTRUM 

PLATES. 

Alongside photographs of spectra, a scale of wave-lengths in 
Angstr5m units is usually given. For the purpose of engraving 
numerals on this scale, the following apparatus has been designed and 
found effective. 

The two essential features are the hot stylus, Fig. i, for ploughing 
a furrow in the gelatine film, and the pantograph, Fig. 2, which guides 
the stylus. The stylus is a steel needle. A, soldered into the end of a 
rod, JBf which is supported by guides at £, permitting it a slight ver- 
tical motion. A small platinum wire, C, has one end wound about 
A near the point, and the other soldered to an insulated wire, Z>, 



MINOR CONTRIBUTIONS AND NOTES 



167 



through which an electric current of about two amperes is introduced, 
heating C nearly to redness, and evolving considerable heat at the 
junction of A and C When the point of the stylus, A^ is lowered 
upon the film of a gelatine plate, the heat is sufficient to soften the film 
in the immediate neighborhood of the point, so that the stylus can 
mould it with ease. A sharp point and thin films are used, so that 
the ridges on each side of the furrow will be small. 

The pantograph is a frame of rigid square brass tubeSi connected 
by vertical hinges at Fy G, H^J. In the arm /K, and perpendicular 
to it, there is introduced a horizontal hinge permitting a vertical motion 
to the index at K. The vertical axis of the hinge, &, is rigidly sup- 




Fig. I 



Fig. 2 



ported by means of a microscope stand at the side of a dividing 
engine, thus supporting the E, F^ G, Hy J system in a horizontal 
plane. When the index, K^ traces any curve in the horizontal plane, 
the stylus at E traces a similar curve, inverted and on a smaller scale, 
the ratio being GF : GJ, 

The plate to be marked is placed longitudinally on the table of the 
dividing engine ; the marking apparatus is lowered by means of the 
microscope stand until the stylus nearly touches the film ; the electric 
current is turned on by means of a key closed by the operator's knee ; 
the stylus is lowered upon the film by a simple device with the left hand ; 
after which the operator with his right hand guides the index, K, over 
the copy, which has been already adjusted to the proper position on a 
subsidiary leaf attached to the dividing engine. With very little 
practice an operator can put from 100 to 150 characters on a plate per 
hour. Olin H. Basquin. 

Northwestern University, 
December, 1894. 



l68 MINOR CONTRIBUTIONS AND NOTES 

ARTHUR COWPER RANYARD, 

It is with deep regret that we record the death of Mr. A. C Ran- 
yard, which occurred at his home in London on December 14, 1894. 

Arthur Cowper Ranyard was born in 1845. ^^ ^i^ early life he 
was much influenced by De Morgan, from whom he received part of 
his mathematical training. In 1868 he took his degree at Cambridge, 
and three years later he was called to the Bar at Lincoln's Inn, 
where he continued to practice his profession until shortly before his 
death. 

But while busied with his legal duties, he found time for much 
work in mathematics and astronomy. From the time of his election as 
a Fellow of the Royal Astronomical Society in his eighteenth year 
until the beginning of his last illness, he was one of its most active and 
faithful members. As Honorary Secretary from 1874 to 1880, and as 
a member of the Council for twenty years, he contributed as few men 
have done to make the influence of the Society what it is to-day. The 
London Mathematical Society was established partly through his efforts, 
and he served it also as an Honorary Secretary. 

His contributions to astrophysics have been numerous and valu- 
able. Few of the Memoirs of the Royal Astronomical Society are so 
frequently consulted as Volume XLI, in which he brought together 
the observations of total eclipses of the Sun up to 1875. '^^i^ is an 
invaluable work to solar physicists. 

After Mr. Proctor's death he undertook to complete the unfinished 
volume of the Old and New Astronomy and to edit Knowledge. In 
the former the chapters added by Mr. Ranyard are the most valuable 
portions of the book. Knowledge^ too, was much improved under his 
editorship, and by his plan of publishing photogravure reproductions 
of the best astronomical photographs, students were for the first time 
enabled to obtain these valuable records. The broad and generous 
nature of the man is shown by the way in which he sacrificed personal 
interests, and devoted himself to the task of carrying to completion the 
work of another. His association in editing De Morgan's Newton is a 
further example of the same spirit. 

His papers dealt with a wide range of subjects. Solar phenomena 
particularly attracted him, and he brought forward much evidence to 
show that prominences are projected into a resisting medium, and that 
a true solar " atmosphere " does not exist. He contended that the gases 



MINOR CONTRIBUTIONS AND NOTES 1 69 

surrounding the Sun show no such increase in density near the photo- 
sphere as they would If the lower layers supported those above them. 
In the forms of certain nebulae he detected remarkable analogies with 
those of solar prominences. The dark areas on long-exposure photo- 
graphs of the Milky Way and other parts of the sky he considered to 
be masses qf absorbing gas between us and the brighter nebulae. He 
devoted much study to the question of the structure of the universe ; his 
treatment of this subject may be found in the Old and New Astronomy, 

As an investigator he did much good work, especially at the solar 
eclipses of 1870, 1S78 and 1882. His very active life left him so little 
opportunity for research that just before his death he was planning to 
establish a large observatory, and to devote his entire time to astro- 
physical investigation. The spectroheliograph which had been used in 
the recent attempts to photograph the corona at the Observatory on 
Mount Etna was to be attached to his reflecting telescope for work on 
the Sun, and a large reflector of novel and ingenious design was to 
be devoted to stellar spectroscopy. But his death occurred while the 
details of these plans were under consideration. 

In his many-sided life his kindly nature and wide sympathies 
brought him many warm friends. They will keenly feel his loss. 

G. E. H. 

THE DESIGN OF ELECTRIC MOTORS FOR CONSTANT 

SPEED. 
In the last April number oi Astronomy and Astro- Physics^ in a paper 
entitled " Electric Controls and Governors for Astronomical Instru- 
ments," the author advocated the replacement, whenever possible in 
laboratory practice,' of the ordinary driving clock of telescopes or 
chronographs by a properly constructed electric motor. Since this 
article was written a good deal of criticism has been directed, partic- 
ularly by astronomers, against the use of motors for such purposes, 
on the ground that they could not be depended upon to maintain the 
requisite constancy of speed. These criticisms have been in the main 
unfair in this, that the performance of the best modern driving clocks 
has been compared with the performance of simply an ordinary com- 
mercial motor, greatly, of course, to the advantage of the former. 

' By " whenever possible " is meant whenever the introduction of electric or mag- 
netic circuits will not be prejudicial to the performance of other apparatus which has 
to be used in connection with, or in proximity to, the source of motive power. 



170 MINOR CONTRIBUTIONS AND NOTES 

The comparison would have been more fair if the governor of the 
driving clock had been entirely removed and an ordinary friction 
brake applied to the driving shaft instead, or the driving weights 
reduced in amount until they would just suffice to overcome the friction 
of the clock and connected mechanism, and therefore drive the latter 
at an approximately uniform rate. Under these conditions, which 
would correspond very closely to the conditions under which an ordi- 
nary commercial motor would work when connected to the same 
mechanism in place of the clock, there is little doubt that the motor 
would soon prove its superiority. Why then should it not be equally 
superior when compared with the governed driving clock, provided it 
also be provided with an equally efficient means of self- regulation ? In 
theory it is possible to automatically regulate the speed of an electric- 
ally driven prime mover with much greater exactness than can be 
obtained with any form of intermittently acting centrifugal governor, 
such as are customarily used, even on the best modern chronographs ; 
for the action of the latter is radically wrong, and can only result in an 
average rate of speed in the driven mechanism. An inertia governor 
of the Siemens form would be far better in principle, but it is unfortu- 
nately rather complicated and liable, because of its excessive sensitive- 
ness, to get out of order. A much simpler and almost equally efficient 
governor is the form of inertia governor which was illustrated and 
described in the paper referred to.' But no matter what form of 
l^overnor is adopted, the usual method of governing is to supply an 
excess of power, which is absorbed by the governor itself to a greater 
or less degree as the load changes ; whereas the correct method is that 
which is adopted in the governor designed by Professor Keeler for the 
driving clock of the Lick telescope, which is itself a simple form of 
electric motor, and supplies a small amount of additional power as the 
load increases and the rate of the clock tends to diminish.' 

It is but a single logical step to the substitution for this small 
governing electric motor (for it is nothing else) of one large enough to 
supply all the power needed to drive the telescope, chronograph drum, 

' A. and A, 13, 265, April, 1894. The general subject of governors, particularly 
in their application to the steam engine, has recently been ably discussed by Mr. 
Conant, of the Westinghouse Co., in the Engineering Maganine for October, 1894. 
In this article Mr. Conant calls particular attention to the great advantages of the 
inertia principle, and illustrates a design of shaft governor which is, in all its general 
features, the same as that described by me in the above article. 



MINOR CONTRIBUTIONS AND NOTES 17 1 

or whatever other piece of apparatus is to be kept in motion, and so 
designed that the pover is supplied by it in just the amount required 
to preserve a uniform speed under all conditions ; in other words, the 
motor itself is made at once the source of power and the governor for 
constant speed. I have already indicated the conditions which should 
be fulfilled in order to secure this self-governing action. These are, 
of course, perfectly well understood by electrical engineers and involve 
no novel points in design, but, in view of the discussion which has 
been raised in regard to the general question of constancy of speed, it 
may be well to re-state them more in detail, and point out the sources 
of variation which they severally and conjointly overcome. 

Given a perfectly constant £. M. F. of the source of electrical energy 
and a steady load, or more generally a constant resistance to be over- 
come (whether that resistance be the work required to drive the appara- 
tus or that required simply to overcome friction, mechanical or mag- 
netic), any motor which is well designed mechanically will, in a very short 
time, settle down to a perfectly constant speed, which will be nearly 
independent of the temperature of the surroundings. These conditions 
are never met with in practice, for in general there is a constant fluctua- 
tion both in the £. M. F. at the terminals of the motor and in the 
resistance to be overcome. These two causes of variation will now be 
considered separately. 

Variation in the load. This is the only or at least the most serious 
cause of variation under the conditions of commercial practice, and a 
motor is therefore usually designed with a view of compensating for it 
alone. Almost perfect compensation may be obtained in several ways,' 
but the two simplest, as well as most efficient, methods that have so far 
been proposed are : first, the method of compound winding, which was 
invented almost simultaneously by Aryton and Perry in England,' by 
Deprez in France,' and by Sprague in America,^ and which has been 
exhaustively tested by Dr. Frdlich ;' and, second, the method proposed 

' For a general risumi of the various methods of regulation see paper by F. B. 
Crocker in the EUciricai World. 

'Aryton and Perry: "Electro Motors and their Goytnaatni,'' Journal of ike 
Soc. of Tel. Eng, and Elec, la, 305. 

3 Regulation de la vitesse du moteur electrique, Comp, Rend, 100, 11 28- 11 62; 
loa, 645. 

*Elec, World, October, 1886. 

^ EUctrotechnische Zeit June, 1885; see also paper by Professor Riicker, Phil, 
Mag. of same date. 



172 MINOR CONTRIBUTIONS AND NOTES 

by Mr. Mordey,' which is simply to construct a shunt motor upon as 
perfect designs as possible, m., to make the resistance of the armature 
very small, the resistance of the shunt very large, the field of the motor 
very powerful and that of armature very weak, and, finally, to properly 
laminate and proportion both armature and field magnets so as to 
reduce eddy currents and self-induction to a minimum. The self- 
regulating power of a motor so constructed, even under very great 
changes of load, is most remarkable. A lo H.-P. motor (Victoria), 
designed and tested by Mr. Mordey, varied less than i per cent, in 
speed when the load was changed from i to lo H.-P., or looo per 
cent. A change of loo per cent, in load, which is far greater than 
would ever occur in cases such as we are considering, would therefore 
change the speed by less than o.i of i per cent., and a change of lo 
per cent, in load by less than o.oi of i per cent.; a much closer regula- 
tion than is accomplished by even the best driving clocks under the 
same conditions. If, then, the source of electromotive force is con- 
stant, as will be the case if the current is furnished from electric light 
mains or from a good dynamo in the laboratory or observatory, a 
motor designed on the above lines will run, under the usual laboratory 
conditions of use, at what is practically a perfectly constant speed. 

But usually the source of current is a storage battery, which, even 
when properly handled, is subject to slight fluctuations of electro- 
motive force (which, however, ought not to amount to more than a few 
per cent.' when the discharge rate is moderate). For this reason we 

> Phil, Mag, January, 1886. 

* When the cells are carefully charged and handled the electromotive force will 
remain almost absolutely constant on closed circuit, if the discharge rate is small, for 
many hours or even days if the capacity of the battery is great enough. In one case 
which I have in mind, in which a battery of four small cells in series was used to 
maintain a constant current through a circuit of wire resistance coils, I found that the 
current, as read by means of a Weston ammeter, could be depended upon to remain 
constant to within i per cent, of its initial value for two or three days ; in one case it 
remained constant to this degree for over a week. To attain such a result it is neces- 
sary : first, that the cells be carefully attended to and charged at a rate never exceeding 
that given by the makers as a maximum rate, and : second, that the discharge rate 
should not exceed ^ the maximum discharge rate. The difficulty which is experienced 
in many cases with storage batteries is due simply to the fact that they are kccorded 
treatment which no good engineer would ever think of giving to a dynamo or engine under 
his care. They are expected to take care of themselves, and perform admirably under 
all conditions of usage, and if they fail to do so the maker of the battery gets the 
blame, or the storage battery itself is condemned as an " unsatisfactory and unreliable 
thing, anyhow." 



MINOR CONTRIBUTIONS AND NOTES 173 

must design the motor so that it will also be self-regulating for slight 
changes in E. M. F. aC its terminals. This problem has been less con- 
sidered in the design of motors than the preceding one, but it is how- 
ever capable of just as satisfactory a solution ; in fact, the first method 
of compensating for changes of load by means of a compound winding 
on the field compensates also for small changes in electromotive force 
at the terminals of the machine. 

To show this let n, be the number of turns per second, N^ the num- 
ber of lines of force, e, the electromotive force at the terminals of the 
motor, E^ the back electromotive force, 4 and i„ the current through the 
armature and shunt field respectively, r«, r, and r^, the resistances of the 
armature, shunt field, and series field, and S, and 5« the number of 
windings on these two fields respectively, then the speed of the motor 
is given by the equation 

C^S,i,-SM ' ^ ^ 

where C is a constant depending only on the construction of the 
armature, and q a quantity depending on the permeability of the mag- 
netic circuit ; but ^, and therefore i„ which is equal to ^->r,, will vary 
with the source of supply. 

Substituting for e its value in terms of i, and r„ and dividing by 
the coefficients of i« in both numerator and denominator, we obtain 



CqS^ 



'•' i,-^i. 



r.-\-r^ 






(») 



Since all the quantities in this expression are constant except i, i« 
and ^, and since this latter quantity may be made constant under con- 
ditions presently to be discussed, it follows that under these conditions 
n will be constant for all values of i, and /« if only 

or when the ratio of the number of shunt to the number of series 
windings is the same as the ratio of the resistance of the shunt to the 
resistance of the armature and series coil together, which is the usual 
rule for compounding to compensate for changes of load; but which, as 

' See Thompson. iZyuMmr^ EUetrU Machinery, p. 6oi. 



174 MINOR CONTRIBUTIONS AND NOTES 

we have just seen, compensates also for changes in electromotive force at 
the terminals. 

In order that the regulation may be perfect we must, as already 
stated, also satisfy the condition ^= constant. But 

^ o.4ir 

^=J=. , ,Z ,_^* (4) 

where /,„, /«, /and A^, A^, A depend simply on the dimensions of the 
magnetic circuit, and are therefore constant, and only /a, and fi^, the 
magnetic permeability of the iron in the field magnet and the armature 
respectively, vary, q is therefore constant when /t, and ii^ are constant ; 
in other words, when the induction through the field and the armature 
has either a very low value, not exceeding 3000 to 4000 cgs. (lines per 
sq. cm.) for annealed wrought iron, or else a very high value, not less 
than 19,000 to 20,000, for the same material. The first condition is 
the best to adopt, for a number of reasons, among which are higher 
efficiency and, in fact, better regulation, because there is less magnetic 
leakage, and relation (4) is more nearly fulfilled for different values of 
/« and is- In fact, if the magnetic density is low, the first method of 
regulation for constant speed under varying load will also give a fair 
although not a perfect regulation for varying electromotive force. For 
in this case we have ^ 

^^ CN~~ CN = CAT • ^5; 

since r^ is very small and i« is also small if the motor is running in 
light load. If we adopt Frdlich's formula for the magnetic circuit 
(which will be nearly true if the field density is low), we have for a 
shunt motor 

^=^«j:p77. (6) 

where N^ is the maximum number of lines which can be forced through 
the circuit, and e' is the diacritical difference of potential ; /. e,y the 
difference of potential which when applied to the terminals of the 
shunt coil will force through the magnetic circuit a number of lines 
equal to \ N„. 

Substituting this value of -A^in (5) we find 

"-CAT-- 
Hence if e is small compared to e' ^ Le./\i the density of field under 



MINOR CONTRIBUTIONS AND NOTES 1 75 

normal working conditions is very low, n will be nearly constant, even 
for considerable changes in e. The regulation, however, in this case is 
not as perfect as that for changes of load, and it is therefore better in 
all cases where great constancy is desired to revert to the compound 
winding, retaining, however, the principal features of construction 
characteristic of the Mordey method, /. ^., a very low resistance of the 
armature (and also of the series field coil), and a very high resistance 
of the shunt coil; because, under these conditions, the number of 
turns in the series field coil becomes very small (see equation (3) ), and 
the efficiency of the motor correspondingly higher. The other essen- 
tial feature of the design is, as already pointed out, a field magnet very 
large and strong in comparison with the armature, but having very low 
density of magnetization. 

The pole pieces should therefore have a less cross-sectional area 
than the cores, instead of a greater, as is usually the case, and the 
resistance of the air-gap should be made very small by adopting the 
Wenstrdm or covered Pacinotti type of armature. The latter should 
also be finely laminated, very carefully balanced and provided with a 
heavy fly-wheel. 

These conditions and essential features of design for a constant 
speed motor were all clearly stated in the previous article, but without 
any explanation of their relative importance, or of the nay in which 
they would bring about the desired result. I hope that the preceding 
rather extended discussion will show that it is perfectly possible and 
practicable to construct motors which will run at a speed constant to 
certainly within o.i per cent., and probably less, under ordinary con 
dition of use, without the intervention of any special governor. Such 
motors will, it is true, be more expensive than the ordinary commer- 
cial ones, and they will be very heavy as compared to their output, 
but they will, on the other hand, have a very high efficiency, both 
electrical and mechanical. Any of the leading manufacturers of elec- 
tric motors would, I am sure, be ready to furnish motors built on 
these lines which they would guarantee to run to within o. i per cent, 
variation in speed under any reasonable fluctuations of load or of £. 
M. F. at the terminals of the machine. 

It is hardly necessary to discuss in detail the many advantages 
which electric motors have over gravity-driven clocks. Perhaps the 
most important are, ist, the perfect control which the observer always 
has over the speed within wide limits of variation ; 2d, the ease with 



176 Af/yOR CONTRIBUTIONS AND NOTES 

which they may be applied directly at the point where the motion is to 
be effected, thus rendering unnecessary any long mechanical connec- 
tions ; and, 3d, the perfect synchronism with which two motors may, 
when necessary, be driven with respect to each other. To effect this 
last result, however, we must use either an automatic governor for each 
motor, or else two similar alternating current motors, driven from a 
common generator. These methods will now be considered very briefly. 

Governors. As we have already stated, we may attain as great 
constancy of rotation as is usually desired with a properly designed 
electric motor, without the use of any special regulator or governor. 
But this is not all. We may, by proper means, control the speed of 
rotation so as to make it absolutely synchronous with the beats of a 
tuning fork, which becomes thus a governor of the motor. 

This can be done by making the fork break the current through 
the armature as many times in each revolution as there a're poles to 
the field magnet. This method, as it was originally proposed and 
used by Deprez' and later by Rayleigh, in his determination of the 
ohm, is only applicable to very small motors, because only a small 
current can be thus interrupted. But it is evident that we do not need 
to interrupt the entire current through the armature, any more than in 
an ordinary driving clock we need to arrest the entire falling weight 
at every revolution of the governor. We need only interrupt a current 
through a very small auxiliary armature, which is attached to and 
revolves with the large armature which furnishes the main portion of 
the power ; or more simply still, interrupt the current through an 
auxiliary winding consisting of a few turns of fine wire disposed side 
by side with the main armature winding, but connected with a separate 
commutator. This is, as I stated in my previous paper, substantially 
the plan which I proposed in 189 1 for driving a revolving mirror in 
synchronism with a tuning fork. The experiments which were begun 
at that time were interrupted by my departure from Clark University, 
and I have never since had an opportunity to resume them. 

The method has, however, been independently developed by Pro- 
fessor Webster, of that University, who has obtained highly satisfactory 
results. "Trials of an experimental machine have shown a regulation 
with an accuracy of one part in ten thousand, and the transmission of 
one-third of a horse-power has been so governed. It is to be hoped that 

'In the former paper I credited this method to Lord Rayleigh, but I have since 
found that it had been previously suggested by Deprez in 1878. See Deprez. 



MINOR CONTRIBUTIONS AND NOTES 177 

the machine, which is to be at once constructed, will furnish still better 
results."' Thomson and Tait, in their treatise on Natural Philosophy, 
suggest, as an ultimate standard of accurate chronometry, "a metallic 
spring, hermetically sealed in an exhausted glass vessel," and state that 
''the time of vibration of such a spring would necessarily be more 
constant from day to day than that of the balance spring of the 
best possible chronometer, disturbed as this is by the train of mechan- 
ism with which it is connected." If this be so then it is possible, by 
using such a spring as a governor of an electric motor in the way just 
described, to secure a more uniform velocity of rotation than is pos- 
sible by any other means. 

The use of a tuning fork or vibrating spring as a governor for the 
motor makes it easy to drive two motors in perfect synchronism with 
each other without any connecting circuit between them, by simply 
using two forks which have been tuned to the same pitch. But if syn- 
chronism of rotation alone is desired, without reference to uniformity 
of rotation, a simpler solution of the problem is found in the use of 
two alternating current synchronous motors driven from the same 
dynamo. The use of alternating current motors instead of direct 
current ones has the further advantage that there is a positive electrical 
gearing, so to speak, between the motor and the driving alternator, 
and the ratio between their speeds is simply the ratio between the 
number of poles on the field magnet of the alternator to the number 
on the field magnet of the motor. If, then, we have a single alterna- 
tor driven at a constant speed (by a self-regulating, continuous current 
motor governed by a fork as above described, for example), we can, 
from it, drive any number of motors also at a constant speed, which 
speed will be any multiple or sub-multiple of the speed of the gener- 
ating alternator. Further, any two of the motors so driven will bear 
a constant phase relationship to each other, which will be in general 
unknown, but which may be varied to any degree at will by rotating 
the field magnet frame of one of the motors ; the resulting change of 
phase being simply the angle through which the field has been rotated. 
This makes it possible to attack, experimentally, certain problems of 
considerable interest in terrestrial physics, of which, heretofore, no direct 
method of solution has been offered. p j^ q Wadsworth 

University op Chicago, 
January, 1895. 

'From a Report of the President and Departments of Clark University for 1893. 
See also EUctncal World, 22,\%\, Sept. 2d, 1893. 



178 MINOR CONTRIBUTIONS AND NOTES 

Schmidt's Theory of the Sun, — In the present number is printed, as 
a clear explanation of an ingenious solar theory which has recently 
received some consideration, an article by Mr. Wilczynski on Schmidt's 
theory of the Sun. According to this theory, the sharpness of the 
Sun's limb, and the enormous change of brightness at that place, are 
not caused by correspondingly abrupt changes in the constitution, 
density, or light-radiating power of the solar matter, but are the result 
of refraction in a non-homogeneous medium. The well-known "fish- 
eye " problem of Maxwell deals with conditions of the same kind, and 
in general the phenomena of refraction in a non-homogeneous medium 
are very complicated and remarkable. Thus the Sun, according to 
the theory, although sharply bounded as seen by the eye, is in reality 
a gaseous ball, whose density diminishes indefinitely and without any 
sudden transitions, with increasing distance from the center. In other 
words, the photosphere is an optical and not a material surface. 

The gaseous sphere considered by Mr. Wilczynski is a purely ideal 
one, and additional interest would have been given to his paper if it 
had been shown that the required conditions are perhaps fulfilled in 
some of the heavenly bodies. Various assumptions as to mass, temper- 
ature, etc., are here necessary, which it is usually impossible to verify, 
but Dr. Knopf has shown in his paper on this subject in A, N, 3199, 
that the conditions in the case of the Sun are well within the bounds of 
probability. Even in the case of Jupiter, circular refraction would occur 
if the atmosphere were only a fraction as dense as that of the Earth. 

But however difficult it may be for present theories to account for 
the tenuity of the solar atmosphere immediately above the photosphere, 
and however readily the same fact may be accounted for by the theory 
of Schmidt, it is certain that the observer who has studied the structure 
of the Sun's surface, and particularly the aspect of the spots and other 
markings as they approach the limb, must feel convinced that these 
forms actually occur at practically the same level, that is, that the 
photosphere is an actual and not an optical surface. Hence it is, no 
doubt, that the theory is apt to be more favorably regarded by mathe- 
maticians than by observers. 

In its application to details, so far as this has been attempted, the 
theory is not very satisfactory. Schmidt's explanation of the promi- 
nences as a result of irregular refraction is rejected even by those who 
regard the general theory as a plausible one. Little better than this is 
Mr. Wilczynski's suggestion that the hydrogen and calcium forms of a 



MINOR CONTRIBUTIONS AND NOTES 179 

prominence, which are so nearly identical as seen in the C and K lines, 
really occur at different levels, and are superposed by the effect of the 
peculiar refraction. In that case the hydrogen prominence would 
not even fall in the same vertical line with the calcium prominence 
which caused it. Here again, the observer is convinced that he is not 
dealing with an optical effect, but that the gaseous forms which he sees 
actually occupy the same part of space. 

J. E. K. 



Reviews. 



PUBLICATIONS OF THE LICK OBSERVATORY. VOLUME III. 

Following close upon the valuable record of Mr. Burnham's double 
star discoveries and measures, comes a third volume from the Lick 
Observatory, with a miscellaneous table of contents. Professors Holden 
and Weinek discuss the lunar photographs obtained with the large 
telescope, and some excellent heliogravure reproductions of the photo- 
graphs accompany the text. Professor Hastings follows with a detailed 
report on his examination of specimens of the crown and flint glass 
used in the 36-inch objective, and Mr. O. H. Tittmann, of the United 
States Coast Survey, describes his comparison of the Lick Observatory 
glass scale "A." But the most interesting and important portion of the 
volume is Part IV, in which Professor Keeler publishes for the first 
time the details of his well-known studies of nebular spectra, including 
his determinations of the motion of the planetary nebulae in the line 
of sight. 

In considering the photographs of the Moon taken at the 
Lick Observatory, it should be borne in mind that they were 
not made with a 36-inch telescope, nor even with the full 
aperture of a 33-inch telescope. It is true that the correcting lens, 
which is attached in front of the 36-inch visual objective for the pur- 
poses of photography, has a clear aperture of 33 inches, but it has 
been found by experiment that the sharpest photographs are obtained 
when this aperture has been reduced by a diaphragm to 8 inches. The 
addition of the third lens reduces the focal length of the telescope 
from 694 inches to 570.2 inches. To all intents and purposes, then, a 
perfectly corrected photographic objective of 8 inches aperture and 
about 48 feet focal length would be fully equal to the great objective 
with its reduced aperture. In fact, the advantage should lie on the 
side of the smaller objective, as in it the loss of light by absorption and 
reflection would be considerably less. Thus if the color correction of 
the 33-inch objective were perfect, and if it were well established that 
a reduction of the aperture to 8 inches tends to increase the sharpness 
of the photographs, it might fairly be considered undesirable to 

x8o 



REVIEIVS l8l 

provide great visual telescopes with photographic correcting lenses of 
large aperture for the purpose of lunar photography. The advantages 
due to the long focal length of the instrument might be retained, and 
the serious task of attaching and detaching the third lens avoided, by 
fixing at the objective-end of the tube a two-lens photographic objec- 
tive of about 8 inches aperture, with a focal length equal to that of the 
visual object-glass. For lunar, solar or planetary photography it 
would be unnecessary to provide a second tube parallel to the large 
one. The saving in labor, weight and expense would be sufficient to 
recommend this plan. Let us hasten to add, however, lest we be again 
called upon to listen to the arguments of those who will not believe in 
great telescopes, that the immense advantages of a large aperture and 
long focus in all visual observations, and in photographic work with 
the spectrograph, are not affected by such considerations. 

At the Lick Observatory the aperture was reduced partly because 
the exposure was inconveniently short with the full 33 inches. But a 
less sensitive plate would have made possible the increase in time of 
exposure, and its finer grain would also have been advantageous. It is 
evident that the difficulty must lie in the large aperture itself. Imper* 
fections m the objective might be supposed to account for this, and 
Professor Holden remarks that "test photographs of stars out of focus 
show that the diluted disk of a bright star is not entirely round, and 
that the light is not spread uniformly over it, as it should be. The 
center of the objective is too short in focus, and a prominent ring 
remains near the outer border.*' These defects are attributed to the 
fact that the final corrections of the lens were made when the tempera- 
ture of the room was below that of the freezing point of water. But 
these defects are not the sole cause of the photographic difficulties. 
The writer has found that better photographs of the Moon are secured 
with a perfectly corrected 12-inch objective when it is diaphragmed 
down to 6 inches than when it is used with full aperture. This in spite 
of the fact that the time of exposure in the former case is four times as 
long as in the latter. On account of the trembling of the image it is evi- 
dent that, other things being equal, a short exposure should give a 
sharper picture than a long one. Hence the half-aperture must oflfer 
sufficient advantages to more than atone for the loss due to increased 
exposure. It seems probable that the difficulty is in large part due to the 
greater effect of atmospheric unsteadiness in the case of the large aper- 
ture. As with good atmospheric conditions a large aperture is far 



1 82 KEVIEIVS 

superior to a sinall one from a visual point of view, it would seem that 
a large aperture should be equally advantageous for photography, if the 
exposure could be reduced to about y^^^ of a second or less. Experiments 
in photographing the Sun would probably render possible the decision 
of this point. It should be remembered, however, that the moments of 
best definition are very brief and infrequent. The eye makes the most 
of them, but few photographs are secured under such advantageous 
conditions. 

We have thought it worth while to dwell upon these considerations 
because they have received little attention in the general discussion of 
the photographs in various astronomical journals. It is probable that 
the question will soon be taken up in comparing the Lick photographs 
with those recently made with the egucUorial coudi of the Paris Observ- 
atory by M. Loewy. 

Lunar photographers will find many other matters of interest in the 
Introduction to Part I. Of the two enlarging lenses used with the 
telescope, that magnifying 8 diameters is found to be unsuitable for use 
under ordinary conditions ; the other magnifies about 5 diameters, and 
is more generally useful. It is remarked, however, *' that we have never 
yet obtained enlargements by this method which were so good as equal 
enlargements in the camera from negatives taken in the principal 
focus. With a telescope like our own, the secret of success seems to be 
to photograph the Moon in the focus on the quickest possible plates, 
with the shortest possible exposures, especially if rapid plates can be 
manufactured which have a fine grain, which seems to be probable. 
Excellent direct 5-fold enlargements of Jupiter have been made, but 
in this case we have short exposures (three seconds) and no appreci- 
able proper motion of the object to deal with. The effect of wind 
on a long-focus telescope is considerable. The photographic focal- 
length of the 36-inch equatorial is 47^ feet. If a direct enlargement 
of 8 diameters is employed the effect of wind tremors on the neg&tive 
is the same as if we were photographing in the principal focus of a 
telescope nearly 400 feet in length. If the telescope is used to make 
direct enlargements (as above described), the guiding of the plate so 
as to avoid the effect of tremors, etc., becomes a matter of extreme 
delicacy." Curiously enough, no attempt has been made at the Lick 
Observatory to screen the telescope from the wind. An adjustable 
canvas curtain, covering the lower part of the slit, would surely be of 
great service in protecting the instrument. 



REVIEIVS 183 

The excellent photographs of the Moon made at the focus of the 
Lick telescope in 1888 do not appear to have been surpassed in sub- 
sequent work, but much time has been spent in making a series of 
exposures with the object of covering the whole history of a lunation. 
The photographs are probably superior to all others hitherto made, 
with the possible exception of those obtained by M. Loewy. Certain 
photographs made by the MM. Henry at the Paris Observatory with 
the 13-inch telescope, and others taken at the Harvard Observatory 
with a similar instrument, are probably but little inferior to the Lick 
plates. In both these cases the image was magnified by enlarging 
lenses attached to the telescope. Professor Holden considers the Mt. 
Hamilton photographs to be as good as can be made with such plates 
as were employed. This opinion is based upon comparisons of the 
diameter of the silver grains (about 0.0002 inches, corresponding to 
o'.07 of arc) with the diameters of some of the small rills and craters 
discovered by Professor Wei nek. The images of some of these craters 
are said to be less than 0.0004 inches in diameter. Hence "these 
measures show that under suitable conditions of illumination, etc., the 
Lick Observatory negatives will depict craters and rills whose absolute 
dimensions are comparable with the size of the grains of the sensitive 
plate/' We are inclined to think that some of the stenographers 
who have hesitated to accept Professor Weinek's discoveries of this 
class will find equal difficulty in adopting this conclusion. The full 
aperture of the 36-inch telescope is hardly sufficient to separate a 
double star of one-tenth of a second, even with the best of seeing and 
a high magnifying power. Professor Holden expects further advances 
in lunar photography when plates of greater sensitiveness and finer 
grain are obtained. But if the negatives now made with an aperture of 
8 inches show craters but little over a tenth of a second in diameter, 
great advances can hardly be hoped to result from work with the same 
telescope. 

The ISO pages devoted to Professor Weinek's discussion of the 
negatives afford sufficient evidence that the photographs are of value 
to the selenographer, and will be of interest to those engaged in lunar 
studies. 

The heliogravure plates from direct enlargements and drawings are 
excellent, and have not been surpassed — if they have been equaled 
— by similar reproductions elsewhere. 

Parts II and III contain the reports of Professor Hastings and 



1 84 REVIEIVS 

Mr. Tittmann. These are good pieces of work, but call for no special 
comment. 

Professor Keeler's memoir on his spectroscopic observations of 
nebulae, which fills the remaining seventy pages of the volume, is a well- 
written account of a most important investigation. With it and the 
contemporaneous investigations of Vogel the era of accurate measure- 
ment in astronomical spectroscopy may fairly be said to open, and as 
a consequence it is certain to rank as a classic in astrophysical litera- 
ture. A preliminary account of a portion of the work here described 
has already been published, but many observations are given here for 
the first time, and the interesting details of the method employed prop- 
erly find a place. 

The paper opens with a valuable risumi of previous observations of 
nebular spectra, written with special reference to determinations of the 
positions of the principal lines. It will be remembered that renewed 
interest was aroused in this subject, in 1887, by Professor Lockyer's 
identification of the chief nebular line as the *' remnant " of the mag- 
nesium fluting at X 5006.4. Professors Liveing and Dewar ascribed the 
fluting to magnesium oxide, and denied that it appears at a lower tem- 
perature than the triplet at b. Dr. and Mrs. Huggins compared the 
chief nebular line with the magnesium fluting, but failed to substantiate 
Professor Lockyer's statements regarding the character and position of 
the former. Professor Lockyer repeated his observations, and suc- 
ceeded in confirming his earlier work* Professor Keeler points out, 
however, that as perfect coincidence of the magnesium and nebular 
lines was noted on both November 27 and February 5, an interval so 
great as 0.46 tenth-meters could not have been visible with the appa- 
ratus employed, as the effect of the Earth's orbital motion between the 
two dates must have caused this displacement of the nebular line. 
After Dr. and Mrs. Huggins had repeated and confirmed their obser- 
vations, they requested Professor Keeler to determine the position of 
the chief nebular line with reference to the edge of the magnesium 
fluting, with the aid of the large spectroscope of the 36-inch telescope, 
and to examine the character of the nebular lines. The spectra of the 
nebulae G, C. 4234 and G, C 4373 were observed with a large Rowland 
grating, and it was at once seen that in each nebula the brightest line 
fell well above the lower edge of the magnesium fluting, thus confirm- 
ing Dr. Muggins's conclusions. The important fact that the distance 
separating the lines was not the same in each showed that one or both 



REVIEWS 185 

of the nebulae had a considerable motion in the line of sight. This 
interesting discovery led to the series of measurements included in the 
following pages. In the winter of 1890-91 the position of the chief 
line in the spectrum of the Orion nebula was accurately determined, 
and the displacement of the Hfi line was measured by direct com- 
parisons with the hydrogen spectrum. The only previous measurements 
of the radial motion of the Orion nebula (other than the early obser- 
vations of Huggins) are those of Maunder, whose results vary from 
51 miles of approach to 11 miles of recession. 

Section II is a description of the apparatus employed, and may be 
passed over without special reference, as an illustrated paper on the 
same subject has already been published.' Section III, in which vari- 
ous experiments made for the purpose of detecting instrumenul errors 
are described, should be consulted by all practical workers with the 
astronomical spectroscope. The light from the nebula and that from 
the comparison spark were made to pursue exactly the same path in 
the spectroscope. The actual aperture of the collimator was reduced 
to the effective aperture by a stop, and an image of the spark was 
formed on the slit by means of a lens of larger angular aperture than 
the collimator. Thus the collimator aperture was completely filled 
with light from the spark. The arrangement is evidently superior to 
that in which a narrow beam from the spark is made to fall upon the 
center of the collimating objective. When a vacuum tube was used, it 
was placed parallel to the slit, and not at right angles to it. With 
proper care in adjustment the former method seems to be preferable. 
The various adjustments of the spectroscope were made in the ordinary 
way, and were then subjected to rigid tests by observations on the 
solar spectrum. Comparisons were made of the coincidence of the 
D lines from a sodium flame held before the slit with those given 
by ft spark between magnesium electrodes containing sodium as an 
impurity. This test was repeated with the b lines, magnesium ribbon 
being burned in front of the slit, and a comparison was also made of 
the position of the D lines in sunlight and the sodium lines from the 
spark. In all cases the coincidence was found to be perfect. Further 
experiments showed that if only one-half the collimator lens were illu- 
minated, or if the collimator and telescope were not exactly focused 
for parallel rays, a small displacement of the lines could be seen. 
Finally, a large number of experiments were made to detect any evi- 

^A,tm4A. 1892* p. 140. 



1 86 REVIEWS 

dences of flexure or of constant errors of any kind. It was found that 
both grating and prism gave the same result. Measures -of the interval 
between the lead line at X 5005.6 and the chief nebular line, made at 
four different position angles 90° apart, agreed within the limits of 
error of the measurements, and no evidence of flexure could be dis- 
covered. Personal bias can play no important part in the observa- 
tions, as the lines in the spectra of the heavenly bodies are continually 
shifting in position on account of the Earth's orbital motion, and the 
interval measured is never twice the same. 

A description of the method of observation follows. Throughout 
the work a Rowland grating was used, and the telescope and collima- 
tor were clamped at a fixed angle of 40**. The error of an observation, 
expressed in wave-lengths, was usually less in the fourth spectrum than 
in the third, but as the latter was somewhat brighter both were used. 
The dispersion in the fourth spectrum was about equal to that of 
twenty-four prisms of 60^, while the third corresponded to fourteen 
prisms. In an observation the nebular line was bisected with the 
micrometer wire, after which the comparison prism was moved into 
place, and the interval between the lines measured. The lead line 
at X 5005.6 was usually employed in the comparison, and for work with 
the second nebular line the double iron line at X4957.6 was used. 
Even in the fourth spectrum the principal line in some nebulte was 
so bright that it could be measured almost as accurately as the metallic 
comparison line. The great focal length of the 36-inch telescope was 
useful in giving comparatively large images of the planetary nebulae. 
The brightness of the spectrum, on accounr of the thickness of the 
great objective, was less than it would have been with a smaller instru- 
ment with thinner objective. 

The positions of comparison lines are discussed in Section V. 
Professor Keeler was fortunate in obtaining from Professor Rowland 
his determination of the wave-lengths of the reference lines employed. 
A very careful measurement of the distance between the lead line and 
the edge of the magnesium fluting gave a result of 1.86 tenth -meters. 
Accurate determinations of the wave-lengths of metallic lines are for- 
tunately somewhat more easily obtainable now than was the case when 
the memoir was written, but the need for such determinations will not 
be completely met until Professor Rowland's Important series of papers 
in The Astrophvsical Journal is concluded. 

Section VI contains examples of the observations, with extracts 



REVIEIVS 187 

from the observer's notebook. The measurements of lines in the 
spectra of the Sun, Moon, Venus, Arcturus, the nebula of Orion and 
other nebulae are all of a high order of excellence. At a time when 
the computed radial velocity of Venus with respect to the Earth was 
+ 7.69 miles per second, the measured velocity came out +^-4 miles 
per second — an agreement so satisfactory as to give confidence in the 
measurements of nebular and stellar motions. Independent determina- 
tions of the motion in the line of sight of the Orion nebula, made in the 
third and fourth order grating spectra, gave values of +11-8 and 
+ 12.4 miles per second respectively. 

Observations of the displacement of lines in the spectra of stars and 
planets were occasionally made during the course of the work on the 
nebulae as a check on the adjustment of the spectroscope. Some of these 
are recorded in Section VII. The dispersion employed permitted a 
displacement due to a motion of two miles per second to be seen, and 
the errors of most of the measurements were considerably less than 
this. In observing Jupiter the lines were seen to be twisted when the 
slit was placed parallel to the equator, on account of the axial rotation 
of the planet. 

We now come to the important work on the nebulae. Section VIII 
opens with an account of the determination of the motion of the 
Orion nebula in the line of sight. In the earlier work on the motions 
of the planetary nebulae the fact that the chief nebular line has no 
known terrestrial counterpart rendered direct measurement of its dis- 
placement impossible. The mean wave-length of this line in the spec- 
tra of ten nebulae distributed through the sky was therefore assumed 
to represent the undisplaced position. On account of the small num- 
ber of objects available, and the lack of uniformity in their distribu- 
tion, this method was replaced as soon as possible by direct measure- 
ment of the displacement of the third line in the spectrum of the 
Orion nebula with respect to the Hfi line given by a Geissler tube. 
The interval between the chief nebular line and the Hfi line was then 
measured, and the wave-length of the undisplaced position of the 
former followed at once. 

The motion of the Orion nebula referred to the Sun was found to 
be -j- 1 1.0 + 0.8 miles per second. This motion is probably in large 
part due to the motion of the solar system in space. No relative 
motion of the different parts of the Orion nebula could be detected. 
A careful examination of the nebular lines showed them to be perfectly 



1 88 REVIEWS 

sharp, with absolutely no indication of the fluted appearance recorded 
by Lockyer. 

Descriptions and measurements of the spectra of the great nebula 
in Andromeda and a large number of planetary nebuls follow. An 
attempt was made to detect evidence of axial rotation in the case of 
G, C. 2 1 02, with negative results. Direct comparisons of the third 
line with the Hfi line were made when the nebular line was bright 
enough for this purpose. 

Section IX brings together the results of the determinations of 
motions in the line of sight, and states the probable errors of the 
measurements. The motions range from +30.1 to —40.2 miles per 
second, and approaching motions preponderate, as most of these neb- 
uls lie in the general direction of the Sun's motion. 

In the three following sections the normal position of the chief 
nebular line, the character of the nebular lines, and comparisons of 
the chief nebular line with the spectrum of magnesium are the sub- 
jects of detailed discussion. It is clearly and decisively shown that 
the chief nebular line is not due to magnesium. 

The origin of the nebular lines and the constitution of the nebulae 
are considered in Section XIII. It is pointed out that no lines in the 
solar spectrum can be identified with the first and second nebular lines. 
Professor Campbell has lately shown that the Ha line is visible as a 
very faint line in many nebulae, but the curious fact of the great rela- 
tive brightness of the fffi and By lines, pointed out by Professor 
Keeler, still remains unexplained. In connection with the discussion 
in this section of the condition of the gases in nebulae, Professor 
Wadsworth's remarks in the January number of this Journal (p. 64) are 
of much interest. 

The memoir concludes with a summary of results, and a note added 
in June, 1894, which epitomizes the investigations of Campbell and 
others, made since the memoir was written in 189 1. An excellent 
plate at the end of the volume illustrates the spectra of three of the 
nebulae observed, as well as the positions of the first and second nebular 
lines. G. E. H. 



Recent Publications. 



A LIST of the titles of recent publications on astrophysical and 
allied subjects will be printed in each number of The Astrophysical 
Journal. In order that these bibliographies may be as complete as 
possible, authors are requested to send copies of their papers to both 
Editors. 

For convenience of reference, the titles are classified in thirteen 
sections. 

I. The Sun. 

Deslandres, H. Etudes des gaz et vapeurs du Soleil. I'Astr. 13, 

455-459. 1894. 
Desl ANDRES, H. Photographs of the Solar Chromosphere. Knowl. 

18, 12, 1895. 
GuiLLAUME, J. Observations du Soleil faites k TObservatoire de Lyon 

pendant le troisi^me trimestre de 1894. C. R. 119, 1186, 1894. 
Maunder, £. W. The New Solar Records. Knowl. 18, 10-12, 1895. 
SiDGREAVES, Rev. W. Notes on Solar Observations at Stonyhurst Col- 
lege Observatory. M. N. 50, 6-12, November, 1894. 
Tacchini, p. Protuberanze solari osservate al Regio Osservatorio del 

Collegio Romano nel 3" trimestre del 1894. Mem. Spettr. Ital. 

83, October, 1894. 
Tacchini, P. Imagini spettroscopiche del bordo solare osservate a 

Catania e Roma nel mesi di settembre, ottobre e novembre, del 1893. 

Tav. CCCX. Mem. Spettr. Ital. 23, October, 1894. 
Tacchini, P. Macchie e facole solari osservate al Regio Osservatorio 

del Collegio Romano nel 3'' trimestre del 1894. Mem. Spettr. Ital. 

83, October, 1894. 
Wonaczeck, a. Anton. Zahlungen von Sonnenflecken. A. N. 137, 6, 

1894. 

3. Stars and Stellar Photometry. 

Backhouse, T. W. Two New Variable Stars. Obs*y 17, 402, Decem- 
ber, 1894. 

Eddie, L. A. Colors and Spectra of One Hundred Southern Stars. 
Jour. B. A. A. 5, 89-98, 1894. 

189 



1 90 RECENT PUBLIC A TIONS 

EsPiN, T. E. Two New Variable Stars and the Variable Stars Es. 873 

and £s. 916. A. N. 136, 397, 1894. 
LiNDEMANN, £. Hellegkeitsmessungen von Z Herculis. A. N. 137, 10, 

1894. 
Parkhurst, Henry M. Notes on Variable Stars, — No. 6. Astr. Jour. 

No. 333, 14, 161-163, January 7, 1895. 
Skinner, Aaron N. New Variable in Hydra, SDM. — 1 4^*2893. Astr. 

Jour. No. 332, Z4, 160, December 21, 1894. 
Smith, C. Michie. Observations of the New Variable Z Herculis. A. 

N. 137, 13. 1894. 
Yendell, Paul S. On Chandler's New Short-Period Variable. Astr. 

Jour. No. 332, 14, 160, December 21, 1894. 

5. Planets, Satellites and their Spectra. 

Denning, W. F. Jupiter. Nat. 51, 227-229, 1895. 

Editors Obs'y. The Recent Transit of Mercury. Obs*y 17, 399, 

December, 1894. 
Editors Obs'y. Mars. Obs*y 17, 401, 1894 ; 18, S3i i^95- 
Elger, T. GwYN. Selenographical Notes. Obs*y 17, 397, 1894; z8, 52, 

1895. 
Krieger, F. N. Mondzeichnungen (Die Ringgebirge Casatus und 

Blancanus). Sirius a8, 1-2, 1895. 
LoEWY and Puiseux. Etudes photographiques sur quelques portions 

de la surface lunaire. C. R. 1x9, 875-881, 1894. 
Peal, S. £. Some Remarks on the Fine Linear Markings seen on Some 

of Professor Weinek*s Enlargements of Lick Photographs of Lunar 

Craters. Jour. B. A. A. 5, 99-101, 1894. 
S&RAPHiMOF, W. Observations des taches sur le disque de Jupiter. 

Bull. Acad. St. Petersburg, No. 2, October, 1894. 
Williams, A. Stanley. Notes on Mars in 1894. Obs*y 17, 391-394, 

December, 1894. 

7. NEBULiB AND THEIR SPECTRA. 

Roberts, Isaac. Photograph of the Nebulae H L 143 and II. 536 

Virginis. M. N. 50, 13, November, 1894. 
Roberts, Isaac. Photographs of the Nebulae H I. 84* h 1442, and H 

II. 344 Comae Berenicis. M. N. 50, 12, November, 1894. 

8. Terrestrial Physics. 

Antoniadi, £. M. The Aurora: of September, 1894. Jour. B. A. A. 
5, 106-107, 1894. 



RECENT PUBLIC A TIONS 1 9 1 

BiGELOW, F. H. Solar Magnetism in Meteorology. Am. Met. Jour, ix, 

319-332, 1895. 
Perry, John. On the Age of the Earth. Nat. 51, 224-227, 1895. 

9. Experimental and Theoretical Physics. 

Carvallo, £. Spectres calorifiques. Ann. Chim. et Phys. (7), 4, 5-79, 

1895. 
Jaumann, G. Zur Kentniss des Ablaufes der Lichtemission. Wted. Ann. 

No. 13, 53, 832-840. 
Ketteler, E. 1st es mdglich, die Erscheinungen der Dispersion des 

Lichtes kUnstlich nachzubilden? Theorie der gegenseitigen Beein- 

flussung von Pendel und Luft. Wied. Ann. No. 13, 53, 823-831. 
KOttgen, Else. Untersuchung der spectralen Zusammensetzung 

verschiedener Lichtquellen. Wied. Ann. No. 13, 53, 793-811. 

1894. 
Nichols, E. L. The Distribution of Energy in the Spectrum of the 

Glow-Lamp. Phys. Rev. a, 260-276, 1895. 
Nichols. Ernest. A Method for the Study of Transmission Spectra 

in the Ultra-Violet. Phys. Rev. a, 298, 1895. 
Paschen, F. Die Dispersion des Fluorits und die Ketteler*sche Theorie 

der Dispersion. Wied. Ann. No. 13, 53, 812-822. 
PiCTET, Raoul. £tude sur le rayonnement aux basses temperatures ; 

applications k la th^rapeutique. Arch, de Gen. 32, 561-573, Decem- 
ber, 1894. 
ViOLLE, J. Sur la temperature de Tare eiectrique. C. R. X19, 949- 

951, 1894. 

11. Photography. 

Jones, Chapman. Color Sensitized Plates. Photo. Times No. i, 26, 
54-55. 1895. 

12. Instruments and Apparatus. 

KOnig. a. Ein neues Spectralphotometer. Wied. Ann. No. 13, 53, 
785-792. 1894. 

Newall, H. F. On a Combination of Prisms for a Stellar Spectroscope. 
Proc. Camb. Phil. Soc. 8, Part 3, 13B, 1894. 

Wadsworth, F. L. O. a New Method of Magnetizing and Astaticizing 
Galvanometer-Needles. Phil. Mag. 482-488. November, 1894. 

Wadsworth, F. L. O. Description of a Very Sensitive Form of Thom- 
son Galvanometer, and Some Methods of Galvanometer Construc- 
tion. Phil. Mag. 553-558, December, 1894. 



192 RECENT PUBLIC A TIONS 

1 3. General Articles, Memoirs and Serial Publications. 

Ames, J. S. A Treatise on Astronomical Spectroscopy, a Translation of 

Die Spectralanalyse der Gestime. (Review.) Phys. Rev. a, 308, 

1895. 
Fowler, A. The Lick Observatory. Nat. 51, 201-203, 1894. 
Gore, J. £. The Construction of the Visible Universe. Knowl. x8, 8-10, 

1895. 
Pickering, £. C. Forty-ninth Annual Report of the Director of the 

Astronomical Observatory of Harvard College for Eleven Months 

ending September 30, 1894. 



THE 

ASTROPHYSICAL JOURNAL 

AN INTERNATIONAL REVIEW OF SPECTROSCOPY 
AND ASTRONOMICAL PHYSICS 



VOLUME I M A R C H I 89 J NUMBER 3 



NOTE ON THE ATMOSPHERIC BANDS IN THE SPEC- 
TRUM OF MARS. 

By William Huggins. 

The question of any part of the bands seen by me in 1862- 
1867 in the spectrum of Mars being really due to the planet's 
atmosphere having been raised by the recent observations of 
Professor Campbell, we have taken the opportunity of Mars and 
the Moon being visible together on the same nights to repeat 
my early observations. 

The method employed by me in 1867 to eliminate the effect 
of the absorption of our own atmosphere from the planet's spec- 
trum was to compare with it on the same night the spectrum of 
the Moon when at a similar or lower altitude {Af. N. 27, 178). 
In my later work on the planets by photography, in 1879, 
another plan was adopted, namely, to take the photographs when 
the twilight was just strong enough to give upon the plate the 
spectrum of the sky close about the planet, without enfeebling 
too much that of the planet itself {Phil. Trans, 1880, p. 687). 

On November 8, 1894, six photographs of the Moon's spec- 
trum and four of the spectrum of Mars were taken with differ- 
ent exposures, giving a range of spectrum from F to S in the 

193 



194 WILLIAM HUGGINS 

ultra-violet. They were afterwards compared by placing a Mars' 
spectrum upon a Moon's spectrum, but in accordance with my 
photographs of 1879 no bands, or other modification of the lunar 
spectrum strong enough to be detected in the photographs, 
could be discovered as peculiar to the planet's spectrum. 

On November 8, 10, and 15 we compared by eye the spec- 
trum of Mars with that of the Moon, and observations of the 
spectrum of Mars when near the meridian were made on Decem- 
ber 15, 18, and 20. Great care and caution are necessary in 
attempting to estimate the relative intensities of faint bands in 
the spectra of objects so different as Mars and the Moon, and 
which, though they can be observed within a few minutes of each 
other, cannot be viewed simultaneously. As the judgment of 
the eye is influenced by brightness and breadth of spectrum, 
care was taken to make the lunar spectrum as narrow and about 
as bright as that of Mars. On these three nights the atmospheric 
bands on both sides of D, to which our attention was almost 
exclusively directed, varied considerably in intensity in the 
Moon's spectrum, but were always estimated by us to be rather 
stronger in the spectrum of Mars. The particular terrestrial 
groups on which our estimations were chiefly based are the nar- 
row band from X5928 to about X5935 and the stronger group 
nearer D at about X5910-5925. We strongly suspected that the 
broad atmospheric group which includes D, and extends from 
about X5885 to X 5905, was rather more distinct in the spectrum of 
Mars, the Moon having a lower altitude at the time. 

These comparisons were repeated several times on each night, 
and Mrs. Huggins' independent observations agreed with my 
own that the bands I have named were always rather more 
easily seen in the planet's spectrum than in that of the Moon, 
even when at a lower altitude. 

Though we reserve our final opinion as to whether there are 
any Martian bands which do not correspond with those of our 
own atmosphere, we think it well to put on record that we have 
little doubt of the existence of a band on the blue side of D, 
beginning at the stronger end at about X5860, and traceable to 



ATMOSPHERIC BANDS IN THE SPECTRUM OF MARS 195 

beyond X5840, which does not appear to be the same as any ter- 
restrial group, and may, therefore, presumably be peculiar to the 
atmosphere of Mars. We took pains to be quite sure that we 
were not deceived by the solar lines at that place in the spectrum, 
which, indeed, are quite different in their relative arrangement. 
The visibility of this band is subject to great variation, depend- 
ing, it may be, on the state of the planet's atmosphere. It was 
distinctly seen on November 10. 

A long series of cloudy nights has unfortunately prevented 
us from making our observations more complete, but we think it 
undesirable to delay longer in stating that the result of our work 
is to leave a strong conviction in our minds that the spectroscope 
does show an absorption which is really due to the atmosphere 
of Mars. 

London, January 6, 1895. 



RECENT RESEARCHES ON THE SPECTRA OF THE 

PLANETS. L« 

By H. C. VOGEL. 

The results which were obtained by the first observers who 
undertook the investigation of the spectra of the planets were 
very contradictory, partly on account of the faintness of the 
spectra, and partly on account of imperfect apparatus. In the 
beginning of the year 1870 I endeavored to throw some light 
on this department of celestial spectroscopy ; by my own obser- 
vations, in the first place, and in the second place by a careful 
and critical comparison of such observations as had already been 
made. My memoir entitled **Untersuchungen fiber die Spectra 
der Planeten," which had been awarded a prize by the Scientific 
Society of Copenhagen, appeared in print in the year 1874.* 
Thanks to the excellent instruments which were at my disposal 
in the private observatory of Herr von Biilow, and to the extreme 
sensitiveness of my eye, I was able to see and to measure so 
many details in the spectra of the principal planets that no 
observations, really subverting my results were to be expected in 
the future. As a matter of fact, neither the great instruments 
of modern times, nor the application of photography — so power- 
ful in the field of celestial spectroscopy — has done more in this 
direction than to confirm the results of these early observations. 

At first sight this seems to be a surprising fact ; it is, how- 
ever, founded in the nature of the circumstances, as I will pro- 
ceed to show here in a somewhat detailed manner. 

In the case of most of the larger planets the image in the 
focal plane of a telescope is a surface of considerable diameter, 
only a small strip of which is cut out by the jaws of the slit. 
With a constant ratio of focal length to aperture, the intensity 
of the image, and therefore the brightness of the spectrum, is 

'Communicated to the Royal Prussian Academy of Sciences; translated from an 
advance proof of the SUtungsberichte sent by the author. 
'Leipzig, W. Engelmann. , 



THE SPECTRA OF THE PLANETS 197 

nearly the same for all sizes of telescopes. It is, indeed, some- 
what less in the case of a large telescope, as the absorption is 
then greater on account of the thicker object-glass. The advan-^ 
tage of the large telescope in the case under consideration is to 
be found in the fact that the spectrum appears broader (higher), 
a condition which is more favorable to the recognition of details. 
In all cases, also, where a cylindrical lens must be used with 
small instruments, in order to obtain the breadth of spectrum 
necessary to make the lines visible, the large telescope is the 
more advantageous with respect to brightness, although not to 
the same extent as in the case of the stars, where the brightness 
of the image for different instruments is nearly proportional to 
the square of the aperture. The only advantage of the large 
telescope with respect to brightness seems to be that larger 
spectroscopes can be attached to it, and (similar construction of 
the apparatus being assumed) the slit opened more widely, 
without changing the purity of the spectrum as compared with 
the smaller instrument. This advantage is again partly annulled 
by the fact that the rays suffer a considerable loss by absorption 
in their passage through the greater thickness of glass, the amount 
of which can be numerically stated only for a particular instrument 
and known varieties of glass. Finally, the larger image in the focal 
plane of a large instrument makes possible the detailed investiga- 
tion of the spectra of difiEerent parts of the planet's surface ; but in 
the above are probably included all the advantages for the inves- 
tigation of planetary spectra which are afforded by a telescope 
of the largest dimensions as compared with one of medium size. 
It is true that in recent times the photographic plate has 
been made sensitive to all parts of the visual spectrum; the 
sensitiveness is, however, by no means uniform, and at present, 
therefore, not much is to be expected from the application of 
photog^phy in the way of securing a permanent impression of 
the less refrangible parts of the spectrum. It is precisely in 
these parts, however, that the characteristic absorption bands of 
the planetary spectra lie. There can be no doubt that when 
photographic processes have been perfected in this direction, it 



198 H. C. VOGEL 

will become possible to determine the positions of the character- 
istic absorption bands in the spectra of the planets more accu- 
rately than by the direct observations which have alone been 
available hitherto. However, the accuracy which has already 
been reached would allow some conclusion to be drawn respect- 
ing the origin of the absorption bands (in particular those of 
Uranus) if we could only succeed in producing a similar absorp- 
tion spectrum experimentally. Although no absorption bands are 
to be expected in the more refrangible parts of planetary spectra, 
the confirmation of this supposition by photographs of the spec- 
trum is not without interest. The same photographs have a still 
higher significance in the case of the spectrum of Uranus, inas- 
much as they disprove in the most emphatic manner the bold 
assumption which was advanced some years ago by Lockyer — 
that the spectrum of Uranus is to be regarded not as an absorp- 
tion, but as an emission spectrum. 

I have lately caused photographs of planetary spectra to be 
taken at the Potsdam Observatory, and the present paper is prin- 
cipally devoted to a discussion of the results. Mr. Hugg^ns, to 
whom we are indebted for the first photographs of planetary 
spectra, has also kindly placed at my disposal the whole of the 
valuable material which he has collected, and with regard to 
which he has hitherto published only some very general state- 
ments.' I have accurately investigated all of Mr. Huggins' spec- 
trograms, and give the results of the investigation after those 
obtained at Potsdam. Finally, I have collected the few known 
observations of planetary spectra which have been made since 
1874, in order to give a complete supplement to my work on 
planetary spectra mentioned in the introduction. 

Data respecting the construction of the two spectrographs, with 
which most of the Potsdam photographs of planetary spectra were 
made, are to be found in my papers on the new star in Auriga* and 
the spectrum of /3 Lyrs.3 These two instruments have been used in 

« PhU. Trans,, 1880, Part II, p. 687. Proe. Roy, Sac., 46, 231. 
• AbkandlMngen dtr K, Pr. Akad, der Wisstnseh,, 1893, S. 8. 
^SUtungsbtrichU dtr K. Pr. Akad. der Wissensck., X894, VI, S. 1 15. 



THE SPECTRA OF THE PLANETS 199 

connection with the thirteen-inch refractor ; the improved spec- 
trometer has been so used since 1893. Photographs of the spec- 
trum of Venus, the brightest of all the planets, were already at 
hand. They were taken with the spectrograph of very high dis- 
persion which I used in connection with the eleven-inch refractor 
of the Observatory for investigating the motions of stars in the 
line of sight. The apparatus is completely described in the 
memoir devoted to the results of this investigation, published by 
the Observatory.' A quite similar instrument, with only a single 
prism, but considerably exceeding in dispersion the instruments 
first mentioned, was sometimes used in connection with the 
eleven-inch refractor for obtaining photographs of the spectra of 
Mars and Jupiter. 

Most of the photographs in the year 1892 were taken by Mr. 
Frost, those of a later date by Mr. Wilsing. For the measure- 
ments, which were all made by myself, a microscope was used 
which has already been described.' On some of the photographs, 
which were taken in the twilight, the sky spectrum is visible on 
each side of the spectrum of the planet, and these plates allow a 
direct comparison of the lines in the two spectra. On other 
plates the spectrum of some neighboring bright star was photo- 
graphed on each side of that of the planet, and on these also a 
direct comparison was possible when the star belonged to 
the second spectral class. If the star was of the first class, at 
least the hydrogen lines could be identified, and they served as 
reference points for further measurements. Most of the plates, 
however, were made without comparison spectra, since the lines 
H and K and the group at G, which cannot easily be mistaken, 
served as reference points from which to determine by measure- 
ment the positions of other lines. 

For determining the wave-length of a line, the distance of 
which from a known line is given in revolutions of the microm- 
eter, I have used the customary and well-known method of 
drawing curves for each spectroscope from numerous measures 

» Bd. VII, I Thcil (Nr. 25). S. 7. 

^PM, d. AUroph. Ods„ Bd. VII, I Theil, S. 31. 



20O H. C. P^OGEL 

of solar spectrograms and computing tables based on the 
curves. 

A description of the apparatus with which Muggins obtained 
his photographs of planetary spectra is to be found in his classical 
memoir **0n the Photographic Spectra of Stars."' 

The length of the spectra is considerably less than that of the 
spectra obtained with the Potsdam photographic refractor. The 
following comparison will give an idea of the relative dispersion 
of the different spectrographs : 

Ufuar Extent of the Spectrum from Y to the mean o/H and K. 
Spectrograph I, - - - 69"" ) In connection with the 

" II,- - -• -16 ) II -inch refractor. 

"Ill, - - 7.0 ) In connection with the 

*' IV, - - - - 8.6 J 13-inch refractor. 

** of Huggins - - 5.3 

Spectrograph III was put together in a provisory manner, 
and in the spring of 1893 was taken apart again. Its prism was 
then used with Spectrograph IV. The collimator and camera 
objectives, both of which are achromatized for the chemically 
active rays, have a somewhat greater focal length than those of 
Spectrograph III. 

In order to obtain wave-lengths from the measures of the 
Huggins photographs, I constructed a curve as before, based on 
the measurement of a number of plates on which the lines of the 
air spectrum are very strongly impressed. 

MERCURY. 

Three photographs of the spectrum of Mercury were obtained 
on March 30, 1892. On one of the plates the spectrum is sur- 
rounded by that of the bright sky, which is of such strength as 
to allow a very reliable comparison with the spectrum of the 
planet. The latter extends from X 487 /a/a to X 380 /a/a, and the 
comparison showed a complete agreement of the spectra. Twenty- 
eight lines could be identified. On the second and third plates the 
daylight spectrum is not perceptible, and on account of the low 

»/%j/. Trans., Part II, 1880. 



THE SPECTRA OF THE PLANETS 



201 



altitude of the planet, the spectrum, in which some fifteen lines can 
be recognized, extends toward the violet only as far as K. A 
good plate was also obtained on April 4, 1892. Sixteen planetary 
^ines can be identified with lines in the accompanying sky spec- 
trum. 

VENUS. 

On each of the dates December 7 and December 29, 1888, 
and January 2, 1889, ^^^ photograph was obtained with Spectro- 
graph I, while three plates were exposed on February 10, 1889. 
The length of exposure varied from 1 5" to 20". All the photo- 
graphs are good. I have measured portions of the best one, 
that of February 10, and have identified the lines with the lines 
of Rowland's atlas. A part of the results of the measurements, 
testifying to the great abundance of lines on the plate, is given 
below. The investigation of the whole spectrum, which required 
the work of several days, showed that over 500 lines in the 
spectrum of Venus between X 460 ^ and X 406 /ui/a are identical 
with lines of the solar spectrum, and that with respect to intensity 
also the agreement of the two spectra is perfect. 



X 

4198.5 
4198.8 
4199-2 
4200.2 
4201.0 
4202.2 
4204.1 



Strong 
Weak 



IT0 



X 

4217.6 



>er, but still dit- 
I tincdy seen as 



Fairly strong line; ^mteliiiM' 4218.8 
Delicate line 4219.5 

Broad line 4220.5 

Very strong 4222.4 

Line (preceding is a line at the 4223.5 



limit of visibility) 
Band 
Line 

Weak line 
Weak line 
Line 
Fine line 



4205.3 

4206.9 

4207.3 

4209.0 

4210.5 

4211.1 

4212.0 

4212.8 

4213.8 Broad line 

4215.7 Strong line 

42 1 6. 1 Line 



Delicate lines 



4224.6 
4225.0 
4225.8 
4227.0 

4227.5 
4230.0 
4231.2 
4233.2 

4235.5 
4236.0 



Two delicate lines, blending 

together 
DifiFuse line 
Line 
Line 
Strong 

Perhaps double 
Somewhat wide 
Fine line 
Broad 
Very broad and \ 

very strong >• Double 

Strong line ) 

Very strong 
Very weak 
Three or four lines, the 

strongest toward the red 
Weak 
Very strong 



202 



H. C VOGEL 



X 

4237.2 
4238.2 
4239.0 
4240.0. 
4241.0 



Very strong equal lines 



Broad; somewhat diffuse 



Lines 



^^^^•7 [Broad 
4243.5 ^ 

4245.5 Very strong line 
4246.2 Weak 
4247.0 I 

4247.5 ' 

4248.7 Weak band 

4249.8 Very fine 

4250.2 Strong line 
4251.0 Strong line 
4253.0 Delicate line 

4254.3 Strong line 
4255.5 Broad band ; weak 
4259.0 Broad band ; weak 

4260.4 Broad strong line 

426 1 .8 Broad band ; weak 
4315.2 Strong 

4317.0 Very weak 

4318.9 Distinct 
4321.0 Broad 
4322.0 Very weak 

4323.5 Band ; perhaps lines 
4325.2 Strong 

4326.0 Very strong 

4327.1 Broad ; weak 
4331.0 \ Broad; weak 



X 

4346.8 Doubtful 

4348.1 Distinct 

4349.1 Perhaps a line 

4351.2 Well visible 

4352.0 Strong line 
4353-0 Very strong line 

4355.2 Broad ; weak 

4358.8 Broad line 

4359.9 Very strong line 

^3^3.3 [weak band 

4364.3 5 

4366.5 Broad and diffuse 
Strong line 
Strong line 
Strong line 
Weak 
Band ; not distinctly resolv- 

4376.1 \ able into lines. 



4367.8 
4370.0 
4371.2 
4373.0 

4374.5 [ : 

.1 ) 



4377-5 ) Partially blended weak lines; 
4379.2 ' perhaps still another line 
4381.0) in the group 
4383.6 Very strong broad line 
4385.1 Broad; somewhat weaker 
than the preceding 

Group of lines; not 
resolvable 



4387.0 
4391.0 
4394.1 
4395.2 
4400.0 
4400.5 
4401.5 
4403.5 
4405.0 
4407.0 I 



Line 
Line 
Very strong line 

^'°^ [ Partially 

I blended 



Line 

Strong line 
Weak line 
Very strong line 



4331.9 \ Fine line 

4337.1 Very strong ) p^^ially 

4337.6 [ Y\M lines C blended 
4338.0 S J 
4339.8 Broad ; very strong 

4340.7 /^7i broad and strong 
4343.5 Broad ; weak 
4344.5 Broad ; weak 

The spectrum of Venus has also been repeatedly photo- 
graphed with the spectrographs having less dispersion. An 



4407.9 >- Lines, not ceruinly resolved 

4408.5 I 

44 1 2. 1 Fine line 

4415.0 Very strong line 



THE SPECTRA OF THE PLANETS 



203 



exposure of three-quarters of a minute was sufBcient to give a 
spectrum which extended far into the ultra-violet. 

A photograph taken by Mr. Huggins in 1879 shows the 
spectrum of Venus together with that of the bright background 
of the sky. More than eighty lines can be distinguished in both 
spectra, and not the slightest anomaly can be found in the 
spectrum of the planet. The photographic spectrum extends from 
X 480 fifi to X 320 fifi. The last portion, above X 328 fifi in the 
spectrum of the planet, and above X 334 ^ in that of the sky, is 
very weak. 

MARS. 

Three photographs of the spectrum of Mars were taken on 
July 27 and 29, 1892, with Spectrograph II. With 10" exposure 
(the first photograph of July 27) the spectrum is somewhat weak, 
although many lines can be seen in it. The best of the three pho- 
tographs is the second one of July 27 (30" exposure); it shows 
many very sharp lines. Between F and K, seventy-five lines could 
be identified with lines in the solar spectrum, and no deviation from 
the solar spectrum of any kind could be detected in this region. 
In order that a judgment may be formed as to the quality of the 
results furnished by the apparatus, I will give here the results of 
my investigation of a small part of the spectrum : 



X 






X 




4199-0 


Broad and strong line 


4260.5 


Broad line 


4202.0 


Strong line 




4265.0 


Weak hand 


4204.0 


Line 




4272.0 


Broad line 


4206.0 


Line 




4274.5 


Broad line 


4210.5 


Line 




4280.5 


Broad line 


4216.0 


Strong line 




4287.0 


Broad line 


4219.5 


Weak line 




4290.0 


Broad line 


4222.3 


Weak line 




4294.2 


Line 


4227.0 


Very strong 


line 


4300.0 


Broad strong line 


4233.5 


Line 




4302.5 


Line 


4236.0 


Strong line 




4306.0 


Line 


4238.0 


Broad diffuse band 


4308.0 


Strong line 
System of lines 


4245.0 j 


1 Broad hand ; 


; diffuse 


4313.0 


4250.0 j 


\ toward the violet 


4315.0 


Line 


4255.5 


Broad line 




43^6. 


Band 



I System 






of lines 



204 H. C. VOGEL 

X X 

4341.0 Line, difiFuse toward the violet 4370.0 Broad, weak 

4344.4 Line 4376.0 Broad band 

4352.2 Strong line 4384*0 Very strong line 

4360.0 Line 

Three excellent photographs were also obtained by Mr. WiU 
sing on November i, 1894, with Spectrograph IV. On the first 
plate, which received 90 seconds exposure, the spectrum extends 
from F to a little above H ; fifty lines can be recognized. On the 
second plate, which was exposed 3", there are forty-five lines, and 
on the third plate, which was exposed 5" and shows the greatest 
amount of detail, seventy Fraunhofer lines can be recognized. 
The spectrum can be followed as far as ^373^ although it 
becomes very weak at X383.6/Li/i. 

Mr. Huggins informs me that he obtained several photo- 
graphs of the spectrum in November 1894, which extend far 
beyond the violet, and which also show no departure whatever 
from the solar spectrum. No photographs had been made by 
him before this time. 

As a supplement to the statements in my memoir respecting 
the observations of the visual spectrum made by Huggins in 
1867, I have to add that Huggins has withdrawn the opinion 
which he expressed on the ground of his early observations; 
namely, that the predominant red color of Mars is due to groups 
of lines in the blue and violet.' The photographs also completely 
remove the doubts which remained after the observations of Hug- 
gins as to whether the lines which he perceived in the more 
refrangible part of the spectrum were special lines characteristic 
of the atmosphere of Mars, or merely the Fraunhofer lines, and 
their decision is in favor of the latter alternative. 

Mr. Maunder, in 1877, ^^^ made observations of the visible 
part of the spectrum of Mars, chiefly for the purpose of discov- 
ering any possible traces of the absorptive effect of its atmos- 
phere, and also with a view to detecting any differences in the spec- 
trum of different parts of the surface.' The spectrum of Mars was 

'ilf. A'. a7, 178. 

•iif.A^.38,34. 



THE SPECTRA OF THE PLANETS 20$ 

therefore compared with that of the Moon, at times when 
both bodies were at the same height above the horizon. The 
position of Mars when the observations were made was, however, 
unfavorable. The altitude was only from 24° to 26°, and it was 
a difBcuLt matter to discriminate between the absorption lines 
produced by our own atmosphere and the lines of similar origin 
in the spectrum of Mars itself. Nevertheless, it appeared from 
the observations that some of these lines were broader and 
stronger in the spectrum of Mars than in that of the Moon. 

A further result was that slight local differences in the spec- 
trum of the surface were perceived, appearing as differences of 
relative intensity in entire regions of the spectrum. 

My early observations, to which I will here refer briefly, 
agreed with the observations of Huggins in showing that Mars 
has an atmosphere of similar constitution to our own, the exist- 
ence of which is revealed by certain groups of lines in the vicinity 
of the D and C lines, and by the telluric groups a and 8. Huggins 
was able to observe the planet when it was in a very favorable 
position, but the case was quite the reverse in my own observa- 
tions, for Mars rose little more than 20^ above the horizon, and 
at this altitude the absorption lines of our atmosphere are quite 
noticeable. It was only by most carefully taking account of this 
circumstance, and by making special observations of the Moon 
and fixed stars for purposes of comparison, that I could be certain 
that the delicate telluric groups of lines were strengthened in the 
spectrum of Mars. 

In 1894 Mr. Campbell observed the spectrum of Mars under 
very favorable atmospheric conditions and when the planet was 
at a great altitude. As he was unable to detect any difference 
between the spectra of Mars and the Moon, when both of these 
heavenly bodies were at the same height, he concluded that the 
existence of an atmosphere on Mars cannot be demonstrated by 
means of the spectroscope.^ The investigations of this zealous 
observer, carried out with the powerful instrumental appliances 
of the Lick Observatory, certainly require consideration, although 

3 Pub, A, S. P. 6y 228, 1894. A. and A. 13, 752. 



206 H. C. VOGEL 

in my opinion they are only entitled to equal footing with the 
earlier observations which have been cited ; for, as I have already 
shown, the advantage of a great telescope in this particular case 
is not so considerable that the observations with smaller instru- 
ments need be entirely set aside. 

Incited by this investigation of Campbell's, I repeated my 
observations during the opposition of Mars last year, but on 
account of unfavorable weather I was unfortunately only once 
able to observe the absorption lines in question. The observa- 
tions were made on November 15, 1894, with Spectrograph IV 
(which can also be used for direct visual observations), attached 
to the thirteen*inch photographic refractor. As this telescope has 
a ratio of aperture to focal length of i : 10, it is considerably more 
efficient with respect to brightness than the Lick telescope. The 
atmospheric conditions were unusually favorable. The height 
of the planet was 43'', that of the Moon 25°. The following 
observations were made : 

Group 8 very distinct in the spectrum of Mars, weak in 

the lunar spectrum. 

" a conspicuous in the spectrum of Mars, difficult 

to see in the lunar spectrum. 

" A 5945 \ very distinct in the spectrum of Mars, equally well 

*' X 5920 ) visible in the lunar spectrum. 

With low dispersion a narrow bright interval or band, some- 
what more refrangible than D, is seen in the spectrum of our 
atmosphere, producing almost the effect of a bright line, although 
it is caused by a vacant space among the fine absorption lines 
which occur in the vicinity of the D lines. This bright band was 
easily seen in the spectrum of Mars, but was scarcely visible in 
the spectrum of the Moon. Hence I can only regard this obser- 
vation as a confirmation of my earlier results. On December 1 2, 
1894, when the atmospheric conditions were again excellent, the 
observations were repeated, with the same instrument, by Messrs. 
Scheiner and Wilsing, who were likewise convinced that the tel- 
luric lines appeared with greater distinctness in the spectrum of 
Mars than in that of the somewhat lower Moon. 



THE SPECTRA OF THE PLANETS 20/ 

Mr. Campbell's paper bears evidence of the circumspection 
with which his observations were made. He justly emphasizes the 
importance of making the breadth of the lunar spectrum the same 
as that of the planet. I may remark here, by the way, that I 
also took all con<ieivabIe precautions in my earlier observations ; 
I not only made the lunar spectrum, and the spectra of such stars 
of the first class as I wished to observe for telluric lines, equal to 
the spectrum of the planet in breadth, but I made the lunar 
spectrum as nearly as possible equal to the spectrum of the 
planet in brightness. In only one point I do not agree with Mr. 
Campbell; namely, in his statement that the investigations of 
ThoUon form a contribution to our knowledge of the spectrum of 
the Earth's atmosphere which is of special importance in this par- 
ticular case. Twenty years ago the absorption lines of our 
atmosphere were already very accurately known ; they were at 
any rate quite sufficiently well determined for the case under 
consideration, where it is a matter of less importance to resolve 
the diffuse bands into lines and to recognize the individual faint 
groups, than to draw conclusions from the general impression 
which is produced by all the absorption bands. Now I believe that 
in the endeavor to go too far into details, Mr. Campbell has 
always employed too high a dispersion in his investigation, 
thereby missing a detail of another kind which is of special 
importance in deciding the question whether there is a difference 
between the spectra of Mars and the Moon. Mr. Campbell also 
attaches especial weight to the observation that the absorption 
lines are not more conspicuous at the limb of the planet than 
they are at the center of the disk. I also have at no time sue* 
ceeded in certainly detecting an increase in the intensity of these 
lines ; but this circumstance is, in my opinion, very naturally 
explained, for the increase toward the limb would be a very 
gradual one, and finally a marked increase of intensity is to be 
expected only at the extreme limb of the planet, in such a nar- 
row strip of the spectrum that fine details would no longer be 
recognizable. 

I have received from Mr. Muggins the following information 



208 H. C. VOGEL 

with respect to the observations on the absorption lines of Mars, 
made by himself and Mrs, Huggins in 1894. On November 8, 
10 and 15, the spectrum of Mars was compared with that of the 
Moon by both Mr. and Mrs. Huggins, and on December 15, 18 
and 20 the spectrum of Mars was observed when the planet was 
nearly on the meridian. In the comparisons with the lunar 
spectrum care was taken to observe the two spectra, which differ 
so greatly in breadth and intensity, under as nearly as possible 
equal conditions. 

On the three days of observation above mentioned the 
intensity of the atmospheric bands near D, which were the prin- 
cipal objects of attention, varied considerably in the lunar 
spectrum ; still both observers independently and always accord- 
antly estimated that the groups of lines on which the comparisons 
were chiefly based — a narrow band at A 593 ftfi and a broader one 
at X 592 /ufi — were always stronger in the spectrum of Mars. In 
like manner the broad atmospheric group which contains the D 
lines (X 5887 to X 5903), was repeatedly seen more distinctly in 
the spectrum of Mars, although the Moon was then at a low 
altitude. 

The observers desire to withhold their decision for the 
present as to whether there are absorption bands in the spectrum 
of Mars which do not correspond with those of our own atmos- 
phere ; but they do feel justified in saying now that they have 
but little doubt of the existence of an absorption band, which lies 
a little on the more refrangible side of D, extending from 
X 586 fifi to X 584 fifi, and which has not yet been recognized as a 
telluric group. The lines which are found in this region in the 
solar spectrum have probably made the decision somewhat diffi- 
cult, but they could hardly cause a serious mistake. The visi- 
bility of this band is subject to changes, which, in the opinion 
of the observers, may depend upon the condition of the atmos- 
phere of the planet. 

I believe that it will be necessary to wait for still further 
observations, perhaps also those made from another point of view, 
before the question can be definitely settled. At the same time. 



THE SPECTRA OF THE PLANETS 209 

I must not allow the fact to pass unmentioned, that the existence 
of an atmosphere around Mars is distinctly indicated in the 
photometric observations of Miiller.' This is in contradiction to 
the earlier view, based on a few observations by ZoUner, that the 
atmosphere of Mars must be extremely tenuous, since Mars, con- 
sidered with reference to its phases, behaves like our own Moon. 
MuUer's observations show that in its photometric behavior Mars 
is intermediate between Mercury and the Moon on one side and 
Jupiter and Venus on the other, and that with respect to density 
its atmosphere is more nearly comparable with that of the Earth. 
We should therefore hardly expect that no evidence whatever of 
the existence of a gaseous envelope would be revealed by the 
spectroscope. 

(Concluded in the next number.) 
> A^. d. Astropk, Ods., Bd. IX, p. 330. 



SOLAR OBSERVATIONS MADE AT THE ROYAL 

OBSERVATORY OF THE ROMAN 

COLLEGE IN 1894. 

By P. Tacchini. 

I HAVE the honor to send you a risumi of the solar observa- 
tions made at the ' Royal Observatory of the Roman College, 
during the year 1894. The following are the results obtained 
for the spots and faculae : 





Number 


HclfttiTC r RQoencjr 


RelatiTe Area 


Number 


t894 


ofd«TS 
of 

OMcrvfttion 










of 


of Spots 


of days 
withoat spots 


of Spots 


of Facttia 


SpotOionpe 
per day 


January 


19 


24.37 





IO6.I 


74.2 


7.3 


February 


20 


19.33 





136.3 V 


65.8 


6.3 


March 


20 


17.51 





48.1 


57.5 


4.8 


5£;' 


20 


22.20 





1 14.8 


63.5 


5.6 


21 


32.29 





1 14.6 


80.0 


6.1 


June 


28 


30.93 





138.2 


86.1 


7.1 


July 


31 


28.58 





X25.6 


63.9 


7.1 


August 


31 


24.39 





93.6 


121.5 


5.7 


September 


27 


25.74 





35.9 


108.5 


6.4 


October 


20 


21.55 





95.3 


69.3 


4.5 


November 


25 


17.12 





39.7 


82.8 


4.6 


December 


21 


18.86 





61.9 


82.2 


4.5 



The spots are therefore decreasing as compared with preced- 
ing series. An examination of the records for 1891, 1892, 1893 
and 1894 shows that from September, 1891, to the end of 1894 
the Sun has never been free from spots ; but the period of great 
area and frequency includes the year 1892, and continues until 
July, 1894, with the maximum about the middle of 1893. Even 
in this last series of observations we have noted the almost com- 
plete absence of extraordinary phenomena in the spot regions, 
as regards both quiescent and eruptive prominences ; the spots 
have almost invariably been seen at the Sun's limb in a state of 

calm. 

210 



SOLAR OBSERVATIONS 
The results for the prominences are as follows : 



211 





Promiacaoes 


1894 


Number 

of days of 

Obienrttkm 


Mean 


Mean 


Mean 




Nombcr 


He«ht 


Extent 


January 


14 


6.00 


37M 


r.6 


Fcbruaiy 


18 


7.17 


37 .4 


2 .6 


March 


18 


8.1 1 


37 .5 


2 .2 


April 
May 


18 


5.00 


38.5 


2 .3 


17 


5.94 


35 .7 


I .7 


June 


26 


6.38 


32 .5 


I .7 


July 


31 


4.71 


36 .8 


I .8 


August 


30 


5.20 


36.3 


I .8 


September 


19 


5.53 


38.5 


2 .2 


October 


16 


4.56 


32 .0 


I .8 




23 


4.64 


38 .8 


X .7 


December 


17 


3.41 


35 .3 


I .8 



There has thus been a progressive diminution in the 
phenomena of the prominences. While in the case of the spots 
characteristic secondary maxima occur in the series, the promi- 
nences have almost invariably indicated a state of relative calm, 
and true metallic eruptions have been lacking. A single large 
prominence was observed on December 24 by my assistant, Dr. 
Palazzo, in the southern hemisphere at latitude 29^.5, on the east 
limb, which at 1 1** 26" was 212' high, and at 1 1** 51" had attained 
an elevation of 290'. Later the prominence commenced to dimin- 
ish in height, and simultaneously moved toward the south. It 
seems probable that this was a prominence floating in the solar 
atmosphere, and that it did not take its rise at the point on the 
limb indicated by the first observations.' 

The rather marked variations in the phenomena of spots do 
not agree with the relatively small variations of the prominences. 
As auroras have been few and faint, I find further confirmation 
for my belief that terrestrial auroras are more closely related to 
the phenomena of the chromosphere than to those of spots. 
This does not affect the general accordance that has been found 
to exist between solar and terrestrial phenomena. 

Rome, January 10, 1895. 

*[See in this connection Mr. F^nyi's paper, p. 212. — Ed.] 



A VERY LARGE PROTUBERANCE 



213 



Gieenwich 
MeuTime 


Obwrved 
Height 


Velodty 

in 

Kllometen 


Remarks 


911,5- 


0" 


124' 




Height measured with the filar microm- 
eter. (Fig. I.) 


10 17 


59 


212 


-rl6 


Height determined from time of passage 
across slit. 


18 


29 


208 






18 


59 


220 






19 


30 


225 






20 


I 


211 






20 


33 


217 














Sketch (Fig. 2) made in this interval. 


10 33 


45 


374 


+131 




34 


3« 


388 


-f222 




35 


20 


405 


-f254 




36 


9 


417 


+ 176 




37 


3 


434 


+231 




41 


33 


519 


) 




Velocity determined from the mean of 








k 


+ 194 


the two transits. 


42 


45 


512 


\ 




Very bright up to a height of 310 . 


48? 




545 






49 


23 


568 


^ 85 


Very faint above 413'. 


52? 


4 


626 


4-278 


Very bright up to 338'; faint cloud some 
distance above. 


55 


21 


661 


+ 128 


Very bright up to 444'; continuous mass 
up to 512"; detached portion floating 
above. (Fig. 3.) 


XI 6 


31 


402 






9 


57 


446 




Upper part ver}' faint. 


II 


20 


464 




Upper part very faint. 


12 


44 


397 




Bright up to 297*. (Fig. 4.) 


24 


51 


358 






? 




312 




Bright up to 209'. (Fig. 5.) 


38 


4 


271 




Bright up to 271'. 


38 


57 


272 




Nothing seen above this height. 


39 


24 


151 






12 9 








Only a small elevation visible. 



Though the times of the observations are given here in seconds, tenths of a 
second were used in calculating the heights. 



214 /. FtNYI 

made, but the form is naturally distorted, as the protuberance 
was ascending rapidly; during the 13" required to complete the 
drawing the height increased 157'. Figs. 3, 4 and 5, however, 
are prepared from very hasty sketches, and have no pretentions 
to fidelity, so far as details are concerned ; they are reproduced 
here to show roughly the general outline and structure of the 
prominence, as well as the process of dissolution. 

Certain peculiarities are revealed in the phenomenon now 
under consideration which are worthy of special remark. From 
9** 15" to 10** 17" the protuberance was of a comparatively 
quiescent nature. It rose in this interval of i** 2" with a mean 
velocity of only 16 kilometers per second. Indeed, in the first 
six transits hardly any ascending motion can be recognized, as 
may be seen in the above table. During the interval lo** 20" 
to 10** 33", when Fig. 2 was drawn, the protuberance suddenly 
began to rise; it rose 157'' in the interval of 13", or with a mean 
velocity of 131*^" per second. From thence it rose with still 
greater rapidity, until at 10** 55" 21* it reached the enormous 
height of 11' i' above the limb, or 0.676 of the Sun's radius. 
The dissolution which followed was as rapid as the development, 
but its progress was less uniform. At 1 1** 39" the height was 
only 157''; at this point the observations were interrupted. At 
1 2^ 9" only a small elevation was found at the same place ; the 
entire protuberance had dissolved. At i^ there was also nothing 
more to be seen. 

Now it is very remarkable that the prominence remained so 
long quiescent, and then ascended so suddenly and violently, 
although in the stage of dissolution it did not sink to approxi- 
mately its former height, but disappeared down to the very level 
of the chromosphere. This singular behavior allows the con- 
jecture that the forces which produced such an extraordinary 
ascent were of a peculiar kind ; acting suddenly upon an already 
existing prominence, they caused a disruption in all directions, 
and brought the whole structure to a speedy end. 

The structure of this prominence was quite the same as that 
of the two great prominences of the 19th and 20th of September, 



A VERY LARGE PROTUBERANCE 2 1 5 

1893, which were referred to above. It consisted merely of 
bright bands or strips, which, collected as it were in a bundle, lay 
approximately in the direction of the Sun's radius. The finer 
threads were here and there torn into small oblong fragments, 
very bright in the middle, with somewhat diffuse edges. This 
appearance was perhaps due to the rapid motion in the line of 
sight. I observed during the transits that the protuberance was 
considerably displaced from the image of the slit, sometimes 
toward the red and sometimes toward the blue. The amount of 
the displacement was measured once with the micrometer, the 
measurement agreeing with eye estimates in showing that the 
motion in the line of sight was approximately the same as that 
of ascent. The displaced image did not have the usual conical 
form, but was seen at times with the displaced edge of the slit 
quite sharply defined. 

The breadth of the protuberance strikingly increased during 
the ascent ; it was perhaps doubled. The dissolution proceeded 
from above downward, as may be clearly seen in sketches 3 to 5 
and in the table. The uppermost parts betrayed the most rapid 
changes ; they became paler and vanished. The greatest height 
measured was that of the cloud-like fragment which is shown in 
the upper part of Fig. 3, enclosed in a rectangular frame. It was 
easily seen during the transits, but when I made the drawing, 
working from below upward, it had already become invisible. It 
was therefore drawn only from memory, but probably had a shape 
and position somewhat like that shown. 

The line A. 6677, which usually appears in such eruptions, was 
in this case invisible. There was also no luminous form on the 
Sun's disk which could be regarded as having any relation to the 
protuberance. 

Kalocsa, Hungary. 



ON THE DISTRIBUTION OF THE STARS AND THE 
DISTANCE OF THE MILKY WAY IN AQUILA 
AND CYGNUS. 

By C. Easton.« 

In comparing the results of his observations of the Milky 
Way with the Durchmusterung maps, the writer believes that he 
has found, in a number of places, remarkable analogies between 
the form of the Milky Way and the general distribution of the 
stars which Argelander has included in his g^eat atlas. Lumi- 
nous and obscure spots have frequently been observed in the Milky 
Way, at exactly the same places where condensations and vacant 
regions are found on the charts. The idea occurred to the writer 
to investigate these relations more thoroughly, together with the 
manner in which they could be made evident, and to publish 
such results as he should be able to obtain. 

It is in the first place necessary, after having discussed the 
results obtained by different observers as to the naked eye aspect 
of the Milky Way, to determine, for certain places in the Milky 
Way, the details whose reality seems to be sufficiently well estab- 
lished. I have therefore divided the zones I wished to study 
into a certain number of parts, in such manner that a relative 
value of the intensity of the galactic light could be assigned to 
each. In this manner it is possible to put into a form suitable 
for numerical treatment the complex character of the luminous 
forms in the Milky Way ; avoiding, at the same time, condensa- 
tions and vacancies of an extremely local character, which would 
have a disturbing influence on the determination of the mean 
values. The number of stars for each half magnitude of brightness 
was then counted in the Durchmusterung, in trapeziums measuring 
I ° in declination and 4" in right ascension. (The counts pub- 
lished by Seeliger, embracing 20" in R. A., were hardly suitable 
for this study.) Star gauges have also been utilized as far as 
possible. Star photographs could also be used for this purpose, 

< Dordrecht, Holland. , 



THE MILKY WAY IN AQUILA AND CYGNUS 



217 



both for checking the drawings of the Milky Way compared with 
the star gauges, and for completing the results, in some measure, 
by the aid of counts made on the photographs themselves. 

With the exception of the (unpublished) map of Julius 
Schmidt, which was inaccessible, the writer has been able to con- 
sult all the known drawings and descriptions of the Milky Way, 
namely : those of Heis, Trouvelot, Klein, Houzeau, Gould (for the 
zone in Aquila), Boeddicker, and Easton, and the unpublished 
drawings of Pannekock. The published series of star counts and 
gauges which were available were those of W. Herschel, those 
contained in Vol. II of the Publications of the Washburn Observ- 
atory^ and those of Celoria,* while Dr. Th. Epstein, of Frankfort, 
has had the kindness to send me the results of his star gauges, 
unfortunately not yet published. I have moreover been able to 
consult the three photographs of regions in Cygnus, taken by Dr. 
Max Wolf with exposures of 3, 11, and 13 hours respectively, which 
were so well reproduced in Knowledge (October and December, 
1891). The valued cooperation of Mr. Pannekock, of Ley den, 
has been of great service to me in the course of these researches. 

These researches have led me, in the first place, to a zone con- 
tained between XVIII 20", XIX 40", 0° and + 6° (Aquila), and 
setting aside the gauges of Herschel and Epstein, which are not 
sufficiently numerous in that region, we arrive at the results sum- 
marized in the following table : 

TABLE I. 



A 

B 
A-B 

C 

D 
C-D 



DWbkm 



XVIII 20, XIX O 

XIX 0, XIX 40 

0% + 6' 



XVIII 20, XIX o 

XIX 0, XIX 40 

+ 3% + 6° 



Area in 
Sq. Degrees 



60 
60 

30 
30 



No. of Stan 
Aiselander 



955 

1035 

-80 

432 

566 

-134 



No. of Stan 
Celoria 



2924 
4476 

-1552 
1 189 
2658 

-1469 



Intensity of 
Galactic Light 



II 

III 

-I 

I 

IV 

-III 



* AfM/. Obs, di Brera, XIII. 



2I8 



C. EASTON 



A zone situated in Cygnus between XX 20, XXI 40, + 40** 
and +55^1 was the object of analogous researches, after it had 
been divided into fourteen parts. In order to obtain more 
decided contrasts the strip contained between 47° and 48^ was 
set aside ; further, on account of the inequality of the areas of the 
divisions, the number of stars was replaced by the stellar density 
per square degree. The value attributed to the intensity of the 
galactic light in division A (Table II) is too small, on account 
of the influence of the adjoining great luminous spot y\ — y Cygni. 
The numbers for the other orders of brightness, inferior to the 
9.0 magnitude, have also been inserted in the table. 

* TABLE II. 



A 
B 
C 
D 
£ 
F 
G 
H 
I 
K 
L 
M 
N 
O 



Dtvwions 



XX 20. XXI o 

+ 40% + 43' 
XXI 0. XXI 40 

+ 40% + 43' 
XX 20. XX 32 

+ 43% + 47' 
XX 32, XX 44 

+ 43% + 47'* 
XX 44, XX 56 

+ 43% + 47' 

XX 56, XXI 8 

+ 43% + 47'' 

XXI 8, XXI 20 

+ 43% + 47' 
XXI 20, XXI 32 

+ 43% + 47'* 
XX 28, XX 52 
+ 48% + 50" 

XX 52, XXI 16 
+ 48% + 50'' 

XXI x6, XXI 40 

+ 48% + 50' 
XX 20. XX 44 

+ 50% + 55" 

XX 44, XXI 8 

+ 50% + 55' 

XXI 8. XXI 32 

+ 50% + 55' 





Stellar Density DM. per iqiure degree 




i-«.5 


6.6-7.0 


71-7.5 


7.6-8.0 


8.X-8.5 


8.6-9.0 


9-X-9.5 


0.49 


O.7X 


X.03 


0.85 


a. 72 


5.89 


20.49 


0.36 


0.18 


0.40 


I. II 


2.19 


5.40 


22.90 


0.24 


0.24 


0.35 


X.06 


1.77 


6.12 


19. 43 


0.35 


0.47 


0.24 


0.59 


2.12 


5.30 


22.38 


0.82 


0.82 


0.94 


0.94 


2.47 


5.54 


22.26 


0.47 


0.59 


1. 18 


X.88 


2-35 


4.47 


25.20 


0.35 


0.24 


0.71 


1.06 


2.12 


7.07 


22.14 


0.71 


0.71 


0.47 


1.30 


2.59 


6.00 


25.79 


0.13 


0.38 


0.88 


0.76 


1. 14 


4.42 


17.80 


0.63 


0.38 


0.25 


1.52 


2.15 


6.31 


13.64 


0.51 


0.63 


0.13 


0.88 


2.78 


6.94 


19.82 


0.28 


0.28 


0.65 


1. 00 


X.75 


3.83 


15.33 


0.42 


0.37 


0.60 


0.65 


1.06 


2.72 


10.67 


0.3. 


0.32 


0.78 


1.06 


1.71 


3.23 


14.68 

1 



lau 

of 

Galactic 
Light 



II 
IV 

III 
IV 
V 

VI 
V 

VI 

III 
II 

IV 

II 

I 
II 



For this zone, we have no systematic enumeration of stars 
down to a magnitude less than 9.5 Argelander, like that of Pro- 



THE MILKY WAY IN AQUILA AND CYGNUS 



219 



fessor Celoria. On the other hand, we have at our disposal eight 
gauges of Sir William Herschel and fourteen of Dr. Epstein. 
The latter cover an area of 1400 square minutes ; the telescopic 
field of Herschel was 15'. It has been possible to complete the 
greater part of the gauges by the aid of counts of stars on 
photographs of the same regions, and by the stars in the Durch- 
musterung maps. Classifying the results of these gauges and 
counts under three heads, on the basis of galactic intensity, we 
obtain the following table : 

TABLE III. 



iBteuhrof 
Galactic Liclit 


Area 1400 square minutes 


Field 15 in 


Afgelaiider 

I-IO? 


Wolf. A. 
i-xx,3? 


Epstein 
i-ia? 


Wolf,B. 
1-13.8 ? 


Herschel 


Feeble 

Mean 

Strong 


9. 

12.7 
18.5 


28. 

SI 


65.5 

85.7 

127.4 


165 . 
297. 
492.6 


24.3 
202. 
340. 



It follows from the preceding table that the distribution of 
stars below the loth and down to about the 15 th magnitude 
agrees sensibly with that of the galactic light, and also with the 
distribution of Argelander's stars. Researches made specially 
for the purpose have, however, shown without doubt that, even 
for the zone in Aquila, the Argelander stars play no important 
part in determining the distribution of luminous forms in the 
Milky Way. 

The general result of these investigations seems to be as 
follows : 

I. In the zones which have been considered, the accumulation 
or sparse distribution of stars whose magnitude is even as high 
as 9.5 corresponds to the greater or less intensity of the galactic 
light. 

In Table II there is only one exception to this correla- 
tion (D-E), and that anomaly is probably explained without 
difficulty by an erroneous evaluation of the mean intensity of the 
galactic light in that region. The fact should, however, not be 
lost sight of, that it is necessary to pay attention principally to 



2^0 



C. EASTON 



the analogy of the contrasts between the neighboring regions, in 
comparing the distribution of 9.5 magnitude stars and the 
intensity of the galactic light. 

2. There is a real correspondence of the general outlines of 
the galactic forms with the distribution of stars of the nth mag- 
nitude (Table I), and with the distribution of stars between the 
loth and about the 15th magnitude (Table III). 

3. Thus, in general, for the zones under consideration, tlu 
faint stars which farm the Milky Way are thickly or sparsely scattered 
in respectively tlu same regions as the stars in tlie last class of Arge- 
laTtder, 





Fig. X. Isophotic lines of galactic light 
in the zone a-T Cygni. 



Fig. 2. Stellar density of 9.1-9 5 mag- 
nitude stars (Argelander). 



4. It follows from the above, that, with a very great degree 
of probability, there is a real connection between the distribution 
of gth and loth magnitude stars and that of the very faint stars 
of the Milky Way ; and that, consequently, the faint or very 
faint stars of the galactic zone are at a distance which does not 
greatly exceed that of gth or lOth magnitude stars. 

If the stars of, say, the 13th to 15th magnitude were really 
at the distances that would be indicated by the theory that, their 
brightness is principally determined by their distance, there 
would be no reason why these stellar groups should have very 
nearly the same apparent distribution, in both galactic longitude 
and latitude, as the groups of 9th and lOth magnitude stars. 



THE MILKY WAY IN AQUILA AND CYGNUS 221 

which would be separated from them by such enormous inter- 
vals. 

It would therefore be necessary to admit that in all cases 
where these stellar condensations appear to have an appreciable 
galactic latitude, the real latitude of groups situated at such 
different distances along the line of sight is, for each of them» 
almost exactly proportional to these distances, — which appears 
to be absurd. 

Although these researches include only a small part of the 
Milky Way, it should not be forgotten that the regions which 
are considered extend almost across its full width ; that they are 
quite widely separated ; and that the region in Cygnus, partic- 
ularly, contains spots of gfreat luminosity (like the spot a — A 
Cygni) bordering on spots which are very dark (the "coal sack" 
between/" Cygni and 6 Hev. Cephei). However, it is unneces- 
sary to say that further researches on the aspect of the Milky 
Way to the naked eye are very desirable in order to extend the 
researches outlined above.* 

' The writer will gladly take upon himself the task of centralizing observations 
which others might be inclined to send him, and of furnishing to observers, willing to 
cooperate with him in this way, any information or material (such as the galactic maps 
of Marth) that would be of service in such observations. 



PRELIMINARY TABLE OF SOLAR SPECTRUM 
WAVE-LENGTHS. III. 

By Henry A. Rowland. 







Intentltj 






Inteaaitj 


Wave-length 


SnbMBoe 


ttUi 

Character 


Ware-length 


Subatanoe 


and 
Character 


4125.384 




000 


4130.804 


Ba 


2 


4125.529 




000 


4x30.847 




000 


4125.615 




00 


4x31.013 




oNd? 


4125.776 


Fc 


3 


4131.151 




000 


4125.850 




I 


4131.271 


Mn 


I 


4126.040 


Fc 


3 


4131.419 




000 


4126.200 


V 


000 


4131.507 


Cr 





4126.344 


Fc 


4 


4131.606 




000 


4126.532 




00 


4131.748 




0000 


4126.673 


Cr 


2 


4131.908 







4I26.79« 




00 


4131.950 




000 


4127.007 




I 


4132.100 


V 


a 


4127.070 


Cr 


00 


4132.235 


Fc 


10 


4127.225 




00 N 


4132.435 




oN 


4127.426 


Cr 





4132.560 


Ba? 





4127.529 


Cc 


00 


4132.690 




3 


4127.689 


Ti 


000 


4132.863 




1 


4127.767 


Fc 


4 


4133.062 


Fc 


4 


4127.957 


Fc 


4 


4133.284 




oooN 


4128.093 




00 


4133.441 




000 


4128.251 


V- 


6d 


4133.510 




00 


4128.461 




00 


4133.625 




00 


4128.543 




00 


4133.755 


Fc 


2 


4128.658 


Cr 


000 


4x33.873 




0000 


4128.747 




00 


4133.965 


Cc 





4128.894 




2 


4134.010 


Fc 


3 


4129.054 




00 N 


4134.159 




oN 


4129.127 




000 N 


4134.345 







4129.337 


Cc. 


3 


4134.492 


Fc? 


3 


4129.476 


Cr 


00 


4134.589 


V-Fc? 


3 


4129.616 




2 


4134.675 




I 


4129.657 




000 


4134.840 


Fc 


5 


4129.760 




00 


4135.050 




ooN 


4129.882 




I 


4135.191 


Mn 





4130.112 







4135.325 




0000 


4130.196 


Fc 


2 


4135.447 




od? 


4130.291 




000 


4135.610 




oNd? 


4130.4OX 




00 


4135.838 




00 


4130.520 




00 


4135.9x5 


Zr 





4130.604 







4136.090 




ooN 



TABLE OF SOLAR SPECTRUM WA VE-LENGTHS 223 







Intensitj 






Inteoutj 


Wave-length 


Subet&noe 


aad 
Chafader 


Wave-lencth 


Snbstanoe 


and 
Chaiaccer 


4136.245 




0000 


4143.200 




00 


4136.445 




00 


4143.316 




000 


4136.527 




000 


4143.430 


Ti 


000 


4136.678 


Fe 


4 


4143.572 


Fc 


4 


4136.890 




ooNd? 


4143.664 




2 


4137^33 




000 


4143.782 






[0000 


4137-156 


Fc 


6 


4143.903 






ooN 


4137.272 




0000 


4144.038 


Fc 




15 


4137.428 


Ti, Mn 


oNd? 


4144.130 






ooN 


4137.567 




2 


4x44.238 






X 


4137.809 


Fc,Cc 


I 


4144.353 






00 


4137.925 




00 


4144.402 




00 


4138.038 




00 


4144.538 




ooN 


4138.134 







4144.674 


Cc? 


oNd? 


4138.288 




ooN 


4144.824 




0000 


4x38.515 




oN 


4144.923 




00 


4138.648 




00 


4145.018 




00 


4138.771 




000 


4145.152 


Cc 





4138.909 




000 


4145.242 




0000 


4139.008 







4145.357 




I 


4139.140 




00 


4145.465 




000 


4139.244 







4145.600 




00 


4139.378 




00 


4145.720 




oN 


4139.524 




00 


4145.914 




iN 


4139-611 




000 


4146.020 




00 


4139.764 




00 


4146.133 







4139.885 




00 


4146.225 


Fc 


3 


4140.089 


Fc 


6 


4146.299 







4140.216 




0000 


4146.544 




00 


4140.316 




000 


4x46.655 




00 


4140.400 


Fc? 





4146.845 




oN 


4140.558 




3 


4146.997 




00 


4140.61 1 




00 


4147.145 




2 


4x40.910 







4147.369 




00 


4140.985 




00 


4147.502 




2 


4141.096 




000 


4147.645 


Mn 


I 


414X.208 


Mn 





4147.677 




000 


414M65 


Fe 


oooN 


4147.836 


Fc 


4 


4141.690 




ON 


4148.021 




000 


4141.809 


La 





4148.133 




00 N 


4142.025 


Fc 


4 


4148.330 




00 


4142.180 




ooN 


4148.416 




00 


4142.330 


Cr 


2 


4148.552 




00 


4142.465 




2 


4x48.660 







4142^562 


Ce 


00 


4148.776 




00 


4142.629 


Cr 


2 


4148.878 




00 


4142.744 




2 


4x48.948 


Mn 





4142.923 




0000 


4149.077 




00 


4143000 




00 


4149.285 


C 





4143.103 







4x49.360 


Zr 


2 



224 



NEJS/RY A. ROWLAND 







Intensity 






lalenaity 


Wave-length 


Substanoe 


voA 
Character 


Ware-length 


SnbManoe 


and 
Character 


4149.533 


Fc 


4 


4155.476 




00 


4149.657 







4155.5^^8 




00 


4149.700 




00 


4155.683 




00 


4X49.857 




000 


4155.802 




00 


4149.923 




2 


4155.875 




000 


4150.056 


Cc 


00 


4156.072 




x 


4150.139 




00 


4156.238 


Nd 





4x50.258 




ooN 


4156.391 


Zr 


I 


4150.4XX 




4 


4156.471 




3 


4150.535 




00 


4156.612 




I 


4150.608 


Co 


I 


4156.764 




00 


4150.706 




00 


4156.831 




x 


4150.865 




ooN 


4156.970 


Fc 


3d? 


4150.964 


Ce 


00 N 


4157.167 


Mn 


ooN 


4151.129 


Zr, Ti 


1 


4157.356 


C? 





4151.225 




00 


4157.398 


C? 


000 


4151.360 


C? 


0000 


4157.586 




00 Nd? 


4151-429 


C? 


000 


4157.738 




00 


4x51.569 


C? 


0000 


4157.948 s 


Fc 


5 


4151-638 


C? 


0000 


4158.171 


C 


00 


4151.726 




00 


4158.242 


C 


00 


4151.826 




00 


4158.428 


C 


00 


415X.925 







4158.538 


C? 





4152.108 


Fe,La 


2 


4158.586 


C? 


00 


4152.242 


Cc? 


I 


4158.700 


Co 


00 


4152.343 


Fc 


3 


4158.959 


Fc 


5 


4152.474 


C 


00 


4159.207 




00 


4152.547 


C? 


00 


4159.353 




5 


4152.687 


C? 


00 


4159.401 




000 


4152.759 


Zr 


00 


4159.562 




00 


4152.927 


Cr, La 





4159.640 




00 


4153.071 




ooN 


4159.805 


Ti 





4153.222 


Cr 


00 


4160.025 




oNd? 


4153.283 




000 


4160.256 







4153.404 


C? 


0000 


4160.408 




00 


4153.542 


Fc 


1 


4160.530 




s 


4153.652 


C? 


00 


4160.722 







4153.776 


Co 





4160.942 







4153.971 


Cr 


1 


4161.089 




ooN 


4154.071 


Fc 


4 


4161.239 




2 


4154.265 




2 


4161.369 


Zr- 


a 


4154.364 




000 


4161.471 




000 


4154.448 







4161.571 


Cr 


00 


4154.539 




00 


4161.682 




4 


4154.667 


Fc 


4 


4161.834 




0000 


4154.824 


C? 


00 


4161.961 


Sr 


I 


4154.976 


Fc 


4 


4162.110 




oooN 


4155.126 




000 


4162.281 




ooN 


4155.213 




00 


4162.454 




ooN 


4155.359 




00 


4162.623 




iN 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 225 



Ware-laifth 



iBteatitj 

■ad 
Chwsctcr 



Wmwlength 



Inteasitj 



4162.825 
4163.068 

4163.144 
4163.281 

4163.449 
4163.516 
4163.642 
4163.818 

4x63.959 
4164.069 
4164.180 
4164.313 
4164^23 
4164.493 
4164.673 
4164.804 

4164.945 
4165.I2I 

4165.277 
4165.328 
4X65.550 
4165.676 
4X65.759 
4165.951 
4166.161 
4166.262 
4166.356 
4166.458 
4166.51 1 
4166.691 
4166.823 
4167.013 
4167.126 
4167.197 
4167.332 
4167.438 
4167.560 

4167.737 
4167.884 
4168.025 

4168.133 
4168.279 
4168.336 
4168.446 
4168.634 
4168.784 

4168.957 
4169.110 
4169.253 
4169.4x1 



Ti. Cr- 



Fc 
Cr 
Cc,- 

Ba 



Ce 



C 

Ni 



iN 

00 
000 
• oooNd? 
00 
000 
od 
4 

000 
00 
00 
000 
o 
o 
00 

o 

000 
o 
00 
3d? 
00 
2 

00 


00 
00 
o 

000 
000 

ooNd? 
ooNd? 


000 
000 
8 

00 

iNd? 
iN 
2 
2 

0000 
000 
0000 
ooN 
2 

00 
2 

00 
00 



4169.499 
4169.628 

4x69.775 
4169.926 
4170.009 
4170.149 
4x70.304 
4x70.372 
4170.506 
4170.646 

4x70.797 

4170.900 

4171.068 

4171.213 

4X7X.433 

4X7X.597 

4171.720 

4X7X.854* 

4172.066 

4172.211 

4172.296 

4x72.447 

4x72.509 

4172.641 

4172.748 

4x72.803 

4172.923 

4x73-045 

4X73-X36 

4x73-309 

4x73-480 

4173.624 

4x73-7x0 

4173-841 

4173-950 

4174095 

4174.240 

4x74-343 
4x74479 
4x74-568 
4174.647 
4174807 
4174-973 
4175-082 
4175-292 
4175-383 
4175-496 
4175-625 
4175.806 
4175.941 



Cr 

Cr 



Fc 
Ti,. 



Cr,La,Ma,Ni,Fe 

Ti,Fc 

Al? 

Fc 



Fc 



Fc 



Fc 



Cr 
Fe 



Cr 
Fc 



00 

00 

iNd? 

2 

00 

00 

00 

00 

00 

00 

00 

000 

4 

4 

oooN 

00 

00 

2 

2 

I 

2 

000 

00 



00 

2 
4 
000 

I 

00 Nd? 

2 

3 

3 

00 

00 

3 



000 

00 

o 

000 

ooN 



4 

iN 

000 

00 

00 

5 
000 



'Probably due to some common impurity of unknown origin. Is it Al? or Si? 
The Gallium line is also in this region, but I have no specimen of Gallium with which 
to determine its exact position. The Fc line is strong enough aloiie to account for it 



226 



HENRY A. ROWLAND 







Inlenaity 






lacenBity 




Subitanoe 


and 
Character 


Wflre-length 


SobMBoe 


•nd 
Character 


4176.074 




X 


4183.OXX 




00 


4x76.222 




000 


4183x69 




1 


4176.427 




00 N 


4183.348 




00 


4176.569 




00 N 


4183.480* 


Zr 


xN 


4176.739 


Fe-Mn 


5 


4x83.6x9 




2N 


4177.034 







4183.787 




ooN 


4177.153 




00 


4183.964 




00 d? 


4177.243 







4184.158 




4 


4177.356 




000 


4184.472 




2 


4x77.495 


Nd 





4184.641 







4177.582 




000 


4184.794 




oooN 


4177.698 


Fc 


3 


4185.058 s 


Fc,Cr 


4 


4177.772 




3 


4185.155 




000 


4177.864 




X 


4185.310 




ooN 


4178.012 




xN 


4185.523 




ooN 


4178.160 




00 


4185.705 




ooN 


4178.223 




2 


4185.807 




00 


4178.402 




ooN 


4185.939 




oNd? 


4178.547 







4186.130 




ooN 


4x78.644 




00 


4186.280 


Ti 


X 


4x78.790 




00 


4186.496 


Cr 


oN 


4178.884 




000 


4x86.622 


C? 


000 


4179.025 




3 


4X86.778 


CcZr 


2N 


4179.106 




0000 


4186.955 


C? 


0000 


4179.2x8 




00 


4187.204 


Fc 


6 


4179.359 




00 


4187.409 


La,C 


00 


4179.408 


Cr, Co 





4187.496 


C 


00 


4179.542 


V,- 


3d? 


4187.615 




000 


4179.743 




00 


4187.747 


Fc 


2 


4179.837 




00 


4x87.878 


Zr 





4179.978 


Zr 





4187.943 


Fe 


5 


4180.159 




000 


4x88.019 




3 


4180.206 


C? 


00 


4188.255 


C 


00 N 


4180.3x4 


C? 


000 


4X88.479 




00 Nd? 


4180.405 


C? 


00 


4188.608 




000 


4180.563 




X 


4188.741 


C 


0000 


4180.732 




00 


4188.894 




4 


4180.839 




000 


4189.138 


c.- 


I 


4180.970 


C 


2N 


4x89.264 




I 


4181.065 


Ti 


000 N 


4x89.490 




000 N 


4181.243 




00 


4189.723 


c,. 


2 


4181.353 







4189.983 


V 


oNd? 


4181.515 




ooN 


4190.147 


Mn 


oN 


4181.708 




X 


4190.287 


Cr 


oN 


4I8I.9I9 


Fc 


5 


4190.397 


C 





4182.136 




2 


4190.552 




00 


4182.376 




00 d? 


4190.687 




00 


4182.548 


Fc 


3 


4190.874 


C,Co 


xNd? 


4182.755 




00 


4191.055 




00 Nd? 


4x82.922 




2 


4191.240 




ooN 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 22; 







Intendtj 






Intentttj 


Wm-length 


SubMance 


■ad 
Character 


Wava.leagth 


SubMnce 


and 
Character 


4x91.321 




000 


4198.221 


Fc 


2 


419M33 


C,Cr 





4x98.295 







419X.595 


Fc 


6 


4x98.402 




4 


4191.843 


Fc 


3 


4x98.494 


Fc 


4 


4192.017 


C 


00 


4x98.584 




00 


4x92.171 


Cr 





4198.680 




00 


4192.258 


La? 


00 


4198.800 


Fc 


3 


4192.356 


La? 


00 


4X98.890 




000 


4192.557 


C 


ooN 


4x99.0x9 


Cr 


oooN 


4192.728 




2N 


4199.267 s 


Zr-Fc 


5 


4192.910 




00 N 


4x99.434 




00 


4193.072 


C 


ooNd? 


4x99.535 




00 


4193.274 




ooNd? 


4199.682 




0000 N 


4193.433 




00 


4X99.83X 




00 


4193.545 


C 


00 


4X99.9OX 




000 


4193.61 1 


C 


00 


4200^45 




iN 


4x93.778 




00 


4200.148 


Fc 


2 


4x93.836 


Cr 





4200.261 


Cr 


00 


4x93.964 


C 





4200.451 




000 


4x94.035 


C 


00 


4200.61 X 


Ni 


X 


4194.246 




oooNd? 


4200.761 




X 


4x94.398 


C 


00 


4200.858 







4X94.47X 


C 


00 


4200.946 


Ti 


X 


4x94.646 







42OX.089 


Fc 


3 


4x94.784 


C 


00 


42OX.227 




00 


4x94.886 


Cr 


I 


4201.402 




00 


4x95.006 


Cr 


I 


4201.486 




00 


4195.X55 


C 


od? 


42OX.585 


Zr 


0000 


4x95.3x8 




000 


4201.733 




00 


4195.492 


Fc 


5 


4201.869 


Fc, Ni-Mn 


X 


4x95.572 


C 


00 


4202.198 s 


Fc 


8 


4x95.684 




X 


4202.5x4 




oNd? 


4x95.785 


Fc-C 


2 


4202.668 




00 


4x95.987 




ooN 


4202.750 




00 


4x96.108 


C 


oN 


4202.9x9 




2 


4x96.372 


Fc 


4 


4203.100 




ON 


4x96.5x6 


C 


00 


4203.287 




IN 


4x96.576 


C 


00 


4203.464 




ooN 


4x96.699 


La 


2 


4203.620 


Ti 


ooN 


4196.837 


Fc 


I 


4203.730 


Cr 


2 


4196.929 


C 


00 


4203.85 X 




0000 


4x97.041 


C 


00 


4203.935 


Ti 


00 


4X97.X53 


C 


000 


4204.x 01 


Fc 


3 


4197.257 s 


C 


2 


4204.X63 


La 


4 


4x97.390 


Cr 





4204.359 


Cr 





4197.5x8 




00 


4204.499 




00 


4197.666 




000 


4204.622 


Cr 





4x97.806 




00 


4204.76 X 




00 


4x97.901 




00 


4204.884 




1 


4x98.053 




0000 


4204,9x6 




2 



228 



HENRY A. ROWLAND 



Wave, length 



Snbstaaoe 



4205.054 
4205.186 
4205.239 
4205.421 

4205.545 
4205.702 
4205.894 
4206.059 
4206.127 
4206.289 
4206.461 
4206.583 

4206.735 
4206.862 
4207.059 
4207.104 
4207.291 
4207.410 
4207.566 
4207.788 
4207.982 
4208.110 
4208.266 
4208.333 
4208.415 
4208.514 
4208.609 
4208.766 

4208.941 
4209.015 

4209.144 
4209.347 
4209.521 
4209.660 
4209.762 
4209.914 
4209.985 
4210.224 

4210.494 
4210.561 
4210.662 
4210.783 
4210.860 
4211.127 
4211.350 
4211.512 

4211.675 
4211.795 
4211.899 
4212.048 



C-Cr 



Mn, C 
Zr- 



Intemity 

And 
Character 





000 




1 




I 




000 


Fc 


iN 




2 




000 N 




00 




000 




oooN 









00 




I 


Fc 


3 


Cr 


00 




0000 


Fc 


3 




000 


Fc 


iN 




00 Nd? 









0000 




00 




000 




00 


Cr 







00 


Fc 


3 




00 


C 


00 


Zr 


I 




ooNd? 


Cr 





C 










Cr 





V 


1 


C 


00 Nd? 


Fc 


4 




3 






C 


00 


C 


000 



3N 

00 N 

oN 

00 

000 

o 

2 



4212.200 

4212.393 
4212.564 
4212.801 
4213.010 
4213.135 
4213.323 
4213.434 
4213.579 
4213.679 
4213.812 
4213.991 
4214.070 
4214.198 
4214.295 
4214.406 
4214.527 
4214.634 
4214.787 
4214.994 

4215.075 
4215.221 

4215.337 
4215.459 
4215.581 s 

4215.703 s 

4215.924 

4215.971 

4216.136 s 

4216.351 
4216.516 
4216.760 
4216.966 
4217.061 
4217.220 

4217.365 

4217.420 
4217.720 

4217.917 
4218.038 
4218.214 
4218.384 
4218.558 
4218.726 
4218.885 
4219.081 

4219.177 
4219.358 
4219.516 
4219.580 



Sobstanoe 



Cr? 
C 

C. Cr 
C? 

C 

Fe 

Zr-C 



C 
C 

Fc 
Sr 



C 

Fc 

Cr 



La, Fc-Cr 



-Zr 



Fe 



00 N 
00 Nd? 
00 

3N 
00 

DON 

oN 
000 
00 
000 

3 

o 

o 

00 

000 

00 

00 

00 

00 

00 

00 



00 

00 

2 

5d? 

00 

000 

I 

3d? 



iN 

000 

000 

00 

I 

I 

5d? 

000 

000 

00 

I 

iNd 

000 

3N 

000 N 

000 N 

I 

4 

3 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 229 







Intenaity 






Intensity 


Wave-length 


Substance 


and 
Character 


Wave-length 


Substance 


and 
Character 


4219.756 




000 


4227.129 






iNd? 


4219.895 




00 


4227.316 






ooN 


4220.064 




00 


4227.474 






I N 


4220.212 




I 


4227.606 


Fc 


4 


4220.330 




00 


4227.822 


Ti 


00 


4220.509 


fe 


3 


4227.918 


Zr 





4220.643 




0000 


4228.103 




iN 


4220.738 


Mn 


000 


4228.265 




000 


4220.813 




00 


4228.475 




00 N 


4220.967 




00 


4228.716 




000 


4221.179 




000 


4228.879 




1 


4221.332 




00 


4229.032 




000 


4221.468 




00 


4229.206 




ooN 


4221.633 




IN 


4229.419 




000 


4221.737 


Cr 





4229.566 







4221.853 




00 


4229.677 


Fc 


2 


4221.975 




000 


4229.926 


Fe 


3^, 


4222.183 




000 


4230.075 




I N 


4222.382 S 


Fc 


5 


4230.267 




ooN 


4222.61 1 




00 N d? 


4230.413 




00 Nd? 


4222.768 







4230.560 


Cr 


000 


4222.890 


Cr 





4230.638 







4223.061 




00 N 


4230.7 M 







4223.256 




IN 


4230.861 




000 


4223.396 







4230.985 




00 


4223.509 




00 


4231.183 


Ni 


4N 


4223.643 




I 


4231.360 




000 N 


4223.738 




I 


4231.573 




oooNd? 


4223.891 







4231.767 


Zr 


I 


4224.060 




00 


4231.846 




I 


4224.138 




000 


4231.997 




0000 


4224.337 


Fc 


4 


4232.111 


Nb? 


I 


4224.463 







4232.201 




oooN 


4224.619 







4232.353 




ooN 


4224.673 


Cr-Fe 


3 


4232.544 




00 


4224.792 


Ti 


00 


4232.618 




00 


4225-020 




. 2N 


4232.765 


V 


00 


4225.206 




000 


4232.887 


Fc 


2 


4225.378 







4233.013 




000 


4225.490 




000 


4233.086 


-V 


iN 


4225.619 


Fe 


3 


4233.328 


Mn-Fe 


4 


4225.874 




I 


4233.404 




00 


4225.970 




000 


4233564 




oooN 


4226.116 


Fe 


2 


4233.772 


Fe 


6 


-4226.239 




000 


4234.171 


Co,V 


oN 


4226.3!^ 






000 


4234.385 




oN 


4226.510 









4234.565 




00 


4226.584 


Fe 




2 


4234.707 


Zr 


oN 


4226.724 


Sr? 







4234.895 




000 N 


4226.904 8 g 


Ca 


\ 20 d? 


4235.054 




000 N 



230 



HENRY A. ROWLAND 







iBtensity 






loteostty 


Wsve-Iengtb 


SuoBtuioc 


and 


Ware-lenctb 


Submnce 


and 

Character 


4235.155 




0000 


4242.167 




000 


4235.298 


Mn 


2 


4242.322 







4235.450 


Mn 


3 


4242.443 







4235.679 




000 


4242.535 




2 


4235.796 




00 


4242.615 




2 


4235.894 







4242.766 




2 


4235.994 







4242.897 


Fe 


2 


4236.112 


Fc 


8 


4243.060 




ooooN 


4236.279 




I 


4243.181 




0000 N 


4236.429 


Ni 


I 


4243.364 




Id? 


4236.540 




00 


4243.518 




I 


4236.714 




ooN 


4243.608 


Fe 


3 


4236.801 




00 


4243.714 




I 


4236.966 




Id 


4243.981 




2 


4237.119 







4244.153 




00 


4237.240 




I 


4244.245 




00 


4237.339 


Fe 


3 


4244.406 







4237.412 




I 


4244.500 




00 


4237.655 




oooN 


4244.568 




00 


4237.836 







4244.717 




0000 


4237.946 




000 


4244.888 




00 


4238.050 




00 


4244.971 




00 


4238.188 


Fc 


3 


4245.104 




0000 


4238.398 




00 


4245.243 




ON 


4238.555 


La 


iNt? 


4245.422 


Fe 


4 


4238.778 




00 N 


4245.520 




2 


4238.918 







4245.671 




0000 


4238.970 


Fe 


5 


4245.772 




ooooN 


4239.107 


Cr 


00 


4245.969 




ooooN 


4239.203 




0000 


4246.071 




00 


4239.304 




000 


4246.180 







4239.525 




2 


4246.251 


Fe 


2 


4239.642 




00 


4246.415 




ooN 


4239.759 


Zr 


00 


4246.577 




oN 


4239.890 


Mn 


3 


4246.729 




000 


4240.014 


Fe 


3 


4246.996 


Y? 


5 


4240.115 




I 


4247.279 




ooN 


4240.246 




000 


4247.464 




I 


4240.359 


Zr 


oN 


4247.591 


Fe 


4 


4240.540 


Fe 


2 


4247.726 







4240.622 




I 


4247.886 







4240.763 




000 


4248.057 




000 


4240.87a 


Cr 


I 


4248.215 




oNd? 


4240.959 




000 


4248.384 


Fe 


2 


4241.127 




00 Nd? 


4248.485 


Ti 


00 


4241.285 


Fe-Zr 


2 


4248.575 




I 


4241.485 




ooN 


4248.696 







4241.682 




ooN 


4248.882 




2N 


4241.866 




ooN 


4249.102 




2N 


4242.002 




000 


4249.272 


Tl 


00 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 23 1 







Intensity 






Intensity 


Wave-length 


Sabatanoe 


and 
Character 


Wave-length 


Subatanoe 


and 
Character 


4249.416 




000 


4258.201 


Zr 





424Q.507 




00 


4258.219 




IN 


4249.648 




IN 


4258.477 


Fe 


2 


4249.797 




2N 


4258.639 




ON 


4249.959 




oooN 


4258.774 


Fe 


2 


4250.205 




000 


4258.885 




IN 


4250.287 S 


Fe 


8 


4259.113 


Fc 


2 


4250.629 




00 


4259.249 







4250.863 







4259.305 







4250.9451s 
4251.071)* 


Fc 


8 


4259.460 




iNd? 




I 


4259.664 




000 N 


4251.491 




od? 


4259.917 




oN 


4251.662 




000 


4260.151 


Fe 


2 


4251.783 


Ti 


00 


4260.282 


Fe 


3d? 


4251.905 


Ti 


00 


4260.494 1 




ON 


4252.043 




000 


4260.640 \ S 


Fe 


10 


4252.218 




ooNd? 


4260.768 J 




ON 


4252.388 




00 


4260.888 




I 


4252.468 


Co 





4260.991 


Ti 





4252.618 




00 N 


4261.162 




000 N 


4252.785 




oN 


4261.376 




2 


4252.917 




IN 


4261.496 


Cr, Mn 





4253.157 




I 


4261.679 


-Cr 


2 


4253.363 




I 


4261.748 


Ti 


00 


4253.522 




00 


4261.891 




2 


4253.696 




00 


4262.086 




I 


4253.888 







4262.142 




I 


4254.063 




I 


4262.287 


Cr 


00 


4254.236 




000 


4262.498 


Cr 





4254.505 s 


Cr 


8 


4262.733 




00 


4254.821 




oooN 


4262.864 







4255.002 




00 


4263.126 




000 


4255.134 


Fe 


2d? 


4263.290 


Ti,Cr 


2 


4255.406 




I 


4263.419 







4255.659 


Fe^Cr 


I 


4263.581 




00 N 


4255.791 




IN 


4263.760 


La 


od? 


4255.993 


Fe 


2N 


4263.996 




DOON 


4256.177 




00 


4264.128 




I 


4256.287 


Ti 





4264.370 


Fe 


3 


4256.366 




I 


4264.431 




00 


4256.469 




00 


4264.615 




I 


4256.575 


Zr 


00 


4264.738 




00 


4256.760 







4264.895 


Fe 


2 


4256.966 







4265.083 




00 


4257.294 




ooN 


4265.238 




000 


4257.517 




00 


4265.418 


Fe 


2 


4257.661 




000 


4265.584 




00 


4257.815 


Mn 


2 


4265.696 







4257977 




000 


4265.832 


Ti 





4258.079 




0000 


4266.081 


Mn 


2 



THE MODERN SPECTROSCOPE. XI. 

SOME NEW DESIGNS OF COMBINED GRATING AND PRIS- 
MATIC SPECTROSCOPES OF THE FIXED-ARM TYPE. AND 
A NEW FORM OF OBJECTIVE PRISM. 

By F. L. O. W A D S W O R T H. 

In the construction of astronomical spectroscopes, especially 
those designed for photographic purposes, rigidity is the most 
important mechanical condition that has to be fulfilled. The 
degree to which this may be satisfied in an instrument attached 
to a given telescope will in general depend on two things : ( i ) 
the size of the spectroscope in comparison with the telescope to 
which it is attached; (2) the design of the instrument, both as 
regards the optical and the mechanical features of construction, 
the latter being dependent upon and conditioned by the former. 

As regards the first condition of rigidity, I have already shown 
that for a resolving power not exceeding 20,000 units, a spectro- 
scope of not more than one-half inch aperture will, if properly 
designed, be equal in optical performance to one of four or five 
or even ten times that aperture, and will, of course, be much less 
expensive and bulky, and much more rigid than a larger instrument. 
For each resolving power there is a minimum efficient aperture, 
determined in the case of the grating by the possible fineness of 
ruling, and in the case of the prism-train by the number of 
prisms which may be conveniently mounted and adjusted. Prac- 
tically this limit, in the case of the grating, is one inch to every 
40,000 units of resolving power (for the first order spectrum), 
and in the case of the prism-train, one inch to from 40,000 to 
50,000 units (60,000 to 75,000, if the source is sufficiently 
bright). For instruments of less than one inch aperture the 
question of optical design as affecting rigidity does not need to 
be considered, for with any of the usual forms of spectroscope, 
a sufficient degree of stiffness may be obtained without unduly 
increasing the weight or bulk of the mounting. When, how- 

232 



THE MODERN SPECTROSCOPE 233 

ever, it becomes necessary to use resolving powers as high as 
75,000 to 100,000, or, as is sometimes desirable, to obtain mod- 
erate resolving power with low dispersion (as in the case of some 
classes of bolometric work), or with a large field (as in spectro- 
heliographic work), apertures of two, three and sometimes four 
inches must be used. In this case it is of the highest impor- 
tance to choose that design which, while efficient optically, will 
permit of the most rigid mechanical construction with the least 
weight and bulk ; particularly if the instrument is to be used in 
measurements of precision, or for prolonged photographic expo- 
sures. 

Of the two general types of instrument, the grating and the 
prismatic, the first has heretofore offered the greater number of 
advantages as respects a rigid construction, for, with the usual 
form of prismatic spectroscope, it is necessary, in order to 
observe different parts of the spectral field under the condition 
of minimum deviation, to vary the angle between the axes of 
the observing and collimating telescopes by an amount equal to 
the change in the angle of deviation for the light of different 
wave-lengths ; a change which, in the case of a prism-train of 
only three dense flint or compound prisms, may amount to 15^ 
or 20°. The necessity for providing for this motion makes it 
impossible to obtain as great stiffness of parts as can be obtained 
with a grating spectroscope, in which any part of the spectrum 
may be brought into the field of the observing telescope by a 
rotation of the grating alone, and the two telescopes may there- 
fore be immovably fixed at a small angle to one another, as in 
Hale's spectroheliograph. But, on the other hand, the prismatic 
form of instrument has, for reasons which have been pointed out,' 
two important advantages over the grating; (i) in the more per- 
fect utilization of the light, and (2) in the more uniform distribu- 
tion of actinic intensity throughout the spectrum; advantages 
which render its use almost imperative when the spectra of very 
faint sources are to be measured or photographed. 

' "General Considerations Respecting the Design of Astronomical Spectroscopes." 
Ap, /. Januar>', 1895, p. 70. 



234 F. L. O. IVADSIVORTH 

For this reason a prism-train is almost invariably used in star 
spectrum work, the very class of work in which, on account of the 
prolonged photographic exposures necessary, rigidity is especially 
desirable. The advantages of a design which has all the advantages 
of a grating instrument in this respect, while still allowing of a ready 
examination of all parts of the spectrum at minimum deviation, 
is therefore evident. In a recent paper, " Fixed-Arm Spectro- 
scopes,"' I have shown how this object may be accomplished 
by the combination of a mirror and prism in such relation that 
the plane of the mirror and the plane bisecting the refracting 
angle of the prism intersect on the axis of rotation of the system. 

The first application of this principle to an astronomical 
spectroscope was in the case of the great spectroscope or spec- 
tro-bolometer of the Smithsonian Astrophysical Observatory, 
which is perhaps the largest spectrometer in existence, the col- 
limator having a focal length of nearly ten meters and the observ- 
ing telescope a focal length of about four meters. The condi- 
tions of the work required in this case that the spectrum should 
be passed slowly and continuously through the field of view at 
minimum deviation, and this necessitated, of course, with the 
usual construction, a regular and continuous movement of the 
arm carrying the observing telescope. It was soon found, as 
might be expected, that no great accuracy of measurement could 
be attained with such an arrangement, as the slightest current of 
air or vibration would cause a movement of the end of the 
arm through several seconds of arc' By the adoption of the 
fixed-arm principle, which has enabled all parts of the instru- 
ment, except the prism and its attached mirror, to be rigidly 
fixed in position, this difficulty has been completely overcome ; 
while the ease and certainty of driving the spectrometer circle 
at a uniform rate has been greatly increased by the removal of 
the great weight of the observing telescope arm and its counter- 
poise. Incidentally also the more or less troublesome minimum 

' Phil, Mag, October, 1894; see also A. and A, December, 1894. 
'A movement of 5' would conespond to a movement of only o"^.i at the end of 
an arm over 4"* in length. 



PLATE XI 




THE MODERN SPECTROSCOPE 235 

deviation attachment for the prism has been done away with. 
Successive settings on a particular line will now differ from one 
another by less than the limit oi reading of the spectrometer 
circle, even when the traffic on the streets surrounding the 
Observatory is at its height. 

The success which has attended the use of this type of 
instrument in the preceding case has led me recently to design 
several forms more particularly suited to smaller sizes of astro- 
nomical spectroscopes, and better adapted for use either as a 
prismatic or as a grating instrument without any change in the 
relative position of the two telescopes. 

The first of these, which is designed for a single prism, either 
simple or compound, is shown in Plate XI. Here the general 
form adopted is that in which the angle between the two tele- 
scopes is 90^. The prism P and the mirror M are mounted 
together on a graduated table which may be segmental, or a 
complete circle as shown by the dotted lines, and the angular 
position read by means of a vernier or microscope A^. The 
mirror may be placed between the prism and the observ- 
ing telescope, or between the prism and the coUimating tele- 
scope, as in Fig. la, the latter arrangement being in some 
resf>ects preferable as giving greater freedom from diffused light 
in the field of the observing telescope. In the preceding design 
it is not possible to replace the prism-train with a grating. 

In order to obtain a form in which this may be done without 
changing the relative positions of the two telescopes, as well as 
to obtain greater resolving power for a given aperture, it is desir- 
able to use one of the double-prism arrangements shown in 
Plate XII, Figs. 2 and 3. In the first of these (Fig. 2) two 
similar prism-mirror systems are used, each mounted on its own 
segmental table, the axes of these tables being connected with 
each other and with the axis of a third graduated circle, N, by 
means of gears, or better by steel cords (as in Fig. 2a), so that 
the two systems revolve in opposite directions with the same 
angular velocity as the graduated circle, which thus serves to 
measure the deflection of the ray which passes at minimum 



236 F. L. O. IVADSIVORTH 

deviation, and falls in the center of the field of the observing 
telescope. By means of the steel cord connections (properly 
designed to remain under a constant tension) equality of angular 
movement can be obtained to within a few seconds of arc, so 
that the multiplication of the number of centers of movement 
introduces no serious difficulty in this respect. At any rate, it 
would seem that as great accuracy could be attained in direct 
measurement as with a train of prisms provided with the usual 
form of automatic minimum deviation attachment, which involves 
a number of sliding joints and links ; while the increased sim- 
plicity and rigidity of construction would make it consider- 
ably superior to the latter in measurement by means of compar- 
ison spectra. As will be seen from the figure, the axes of the 
two telescopes are in this case inclined to one another at a small 
angle (which may be made even less than shown, if desired, by 
altering the angle between the prisms and their attached mirrors), 
and intersect at the axis of rotation of the graduated circle. 
Therefore by removing the two prisms and placing a grating at 
the center of A^, the instrument may be converted into a grating 
spectroscope without any change of any of the parts with respect 
to each other. The grating may of course be left permanently 
in position, as it does not in any way interfere with the use of 
the prisms. As in the former case compound prisms may be 
used instead of simple ones. 

In Fig. 3 is shown a somewhat different form of fixed-arm 
double-prism spectroscope with but a single mirror and a single 
circle, which has already been briefly described in a previous 
paper.' In this case the mirror M \s mounted on^an arm which 
is connected permanently to the graduated circle of the spec- 
trometer, and the two prisms (in this figure compound prisms 
are shown, but simple ones may equally well be used, as shown 
in the figure in the Philosophical Magazitiey are mounted together 
on a table, which is connected to the spectrometer table by an 
ordinary minimum deviation attachment (not shown in the 

'"An Improved Form of Littrow Spectroscope." Phil. Mag, July, 1894. 
»/*iV/., Fig. 4. 



PLATE XII 





FigR^ 



Pig a 



t^ j- 






— ^ii^T 


^^^ . IP 





THE MODERN SPECTROSCOPE 237 

figure), so as to revolve at half the angular velocity of the latter. 
To use this instrument as a grating spectroscope, it is simply 
necessary to remove the two "prisms and the minimum deviation 
attachment and place a grating on the same table, which is now 
clamf>ed directly to the spectrometer circle. To more perfectly 
utilize the entire surface of the objectives of both coUimating 
and observing telescopes, the axis of the latter should be dis- 
placed laterally until it intersects the axis of the former near the 
center of rotation (as shown by the dotted line, Fig. 3). This 
is easily accomplished, without sacrifice of rigidity, by slotting 
the braced support for the observing telescope, as indicated in 

Fig. 3- 

In the preceding forms, the general object of the design has 
been to produce a fixed-arm spectroscope of as compact a form 
as possible, and one in which the general arrangement is such 
that either a grating or prism-train could be used without the 
necessity of changing the relative position of the different 
parts. For this reason the position of the mirror or mirrors 
with respect to the prisms has been so chosen that the axes of 
the observing and coUimating telescopes make an angle, j9, of 
nearly 180^ with one another. But one great advantage of this 
general type of instrument is that by simply altering the mirror 
with respect to the prism, any value of j8, from o to 360** may be 
obtained, with either one or two or more prisms, according to the 
conditions of use. In the spectro-bolometer of the Astrophys- 
ical Observatory, for example, it was found to be more conven- 
ient to use an angle j9=o, or in other words, to adopt the direct- 
vision form of instrument' because of the arrangement of piers 
in the laboratory. Similarly in the case of compound spectro- 
scopes, it might be more convenient in some cases to use instead 
of one of the forms shown, either the direct-vision form, or 
one in which the observing telescope was below, instead of above 
the collimator, and the angle jS is therefore between 1 80^ and 
360°. The entire flexibility of this type of spectroscope in this 
respect is one of its important advantages. 

' Cue 4, Fig. 7, "FUed-Ann Spectroccopes," p. 348. 



238 F, L. O, WADSWORTH 

It would not in general be advisable to use more than two 
prism-mirror combinations, both because of the increased loss of 
light by reflection, and the increased expense of the optical sur- 
faces, and because of the increased difficulty of mounting and 
adjusting the systems with reference to each other. But we may 
obtain any number of transmissions through the prism-train, and 
therefore a resolving power equivalent to any number of prisms, by 
the use of double total reflection prisms, which are fixed at the 
end of the train, just as in Young's modification of the origi- 
nal Littrow form ; except that in this case these reflection 
prisms are fixed and not movable with the train — a considerable 
advantage. The type of instrument is, however, adapted more 
particularly to spectroscopes where only a moderate resolving 
power is required, because the rigidity of the construction makes 
it possible to use almost any aperture that we please, and there- 
fore allows the desired result to be attained by the use of one or 
not more than two prisms. Although, as I have previously shown, 
there is very little increased loss of light by reflection for any 
number of prisms above three, there is still a decided gain, 
amounting to nearly 40 per cent, of the total loss, in using only 
one instead of three ; a gain which is well worth considering in the 
case of very faint star spectra, if it can be attained, as in this case, 
without any sacrifice of rigidity, even though the cost be very 
considerably increased. The only remaining difficulty in the 
use of a very large aperture is in obtaining a prism of sufficient 
homogeneity of material, but this difficulty may be overcome 
in a manner which will be discussed in the section devoted to the 
consideration of objective prisms. 

When considerations of resolving power or dispersion demand 
the use of at least three prisms the use of the forms of the 
fixed-arm type so far described becomes impracticable (except, 
perhaps, in the case of the modified multiple transmission form 
already suggested) because of the multiplication of reflections. 
There is, however, another form which, although it has not 
so far been considered to any great degree in the design of 
astronomical spectroscopes, has decided advantages for the 



THE MODERN SPECTROSCOPE 239 

purpose. This is the Littrow type of instrument which, as has 
been recently shown, is only one of the general class of fixed-arm 
spectroscopes, and may therefore be very properly considered in 
this connection. It has, of course, all the advantages of the 
class to which it belongs, as regards rigidity, and the further 
special advantage, which peculiarly adapts it to high resolving 
power — that of requiring (in its simple form) only two reflect- 
ors, one of them quite small, for any number of prisms. 

Its principal disadvantage and the methods of overcoming it 
have already been discussed in the paper in the Philosophical 
Magazine already referred to. In addition to its getting rid of 
the general illumination of the field, the' concave mirror form 
there described has two other features of construction which 
are specially advantageous for astronotnical spectroscopes : 
(i) the concentration of nearly the whole weight of the 
instrument near the slit end, and (2) the position of the reflect- 
ing prism between the slit and the collimator instead of between 
the latter and the eyepiece. The first feature adds greatly to 
the rigidity of the instrument; the second makes the line of 
collimation at right angles to the axis of the telescope, and thus 
allows the main part of the spectroscope to be permanently 
attached to the latter without interfering in any way with its 
independent use. 

In Plate XIII, Fig. 4, is shown an instrument designed on the 
above lines. Jn the figure s is the slit ; a, the small total reflec- 
tion prism which reflects the light from the slit to the coUimating 
mirror d; P, a Littrow prism-train (which may be replaced by a 
grating if desired) of three or more prisms, simple or compound, 
with the usual automatic minimum deviation mounting; and /is 
the observing eyepiece, which may be replaced by a plate holder. 
It will be seen that the latter is so close to both the finder, F, 
and the small telescope, /, which serves to keep the image on the 
slit, that the observer may readily use either of these without 
moving from his position at the eyepiece of the spectroscope. 
Further, by simply removing the box which carries the slit tube 
and its adjustments together with the reflector, ^, and eyepiece. 



240 F. L. O. IVADSIVORTH 

f, the telescope may be used for visual or micrometric observa- 
tions without disturbing the main part of the spectroscope. 

Fig. 5 shows a somewhat similar design of a Littrow grating 
or prismatic spectroscope which is still more compact and rigid. 
Here the light reflected from the slit by the first right-angled 
prism, a, falls upon a second, a^, by which it is reflected along the 
side of the telescope tube to the collimator at b, from which it 
passes to the grating (or prism-train) at G, thence back again to 
b and finally to the observing eyepiece/ Instead of the con- 
cave mirror form we may use the modified Littrow form of Young 
and Lockyer, with separate collimating and observing telescopes 
^, ^', as indicated by the dotted lines. 

This general form of construction will, it is observed, admit 
of the use of a collim'ator and observing telescope of any focal 
length we please (as great, in fact, as the focal length of the large 
telescope), without sacrifice of rigidity and without any objection- 
able increase in bulk. As in the preceding form, all parts of the 
spectroscope save the small portion which carries the slit tube 
may be left permanently in position without interfering in any 
way with the ordinary use of the telescope, an advantage of very 
considerable moment in the case of large instruments, on account 
of the saving of the time and trouble involved both in attaching 
and detaching the instrument itself and in the readjustment of the 
two sets of balance weights on the telescope tube. The only 
objection which can be raised against this or the preceding form 
is the use of the reflectors or total reflection prisms a^ /z„ in the 
path of the rays from the slit to the collimator. 

In general, three objections may be raised against the use of 
reflectors in a spectroscope train. 

1. The increased loss of light due to the additional reflections 
and, in case reflecting prisms are used, the additional absorption 
of the latter. 

2. The impairment of definition due to the increased number 
of optical surfaces and to the distortion of those surfaces due to 
changes of temperature, etc. 

3. Changes in adjustment due to accidental disturbances of 




I 




THE MODERN SPECTROSCOPE 24 1 

the position of the reflectors by reason of temperature changes 
in the mountings, vibration, etc. 

As all of the forms so far described in this paper have 
involved the use of one or more reflecting surfaces, it is of 
importance to consider these objections somewhat in detail and 
to show that none of them affect to any sensible degree the 
efficiency of the spectroscope for either visual or photographic 
or holographic purposes. 

As regards the flrst, the effect of two reflections in diminish- 
ing the brightness of the spectral image is insignificant in com- 
parison with the effect of the prism-train or grating. A well- 
silvered surface or a total reflection prism will reflect 97 to 98 
per cent of the incident light, and the loss due to the two reflec- 
tions must therefore amount to only about 5 per cent, of the 
reflected beam, while the loss by reflection in a single prism 
amounts to about 18 per cent., or over three and a half times as 
much. The proportionate amount of light lost by diffusion and 
multiplication of spectra in the case of the grating is of course 
still greater. We may therefore employ at least two reflections 
for every prism in the spectroscope train without seriously 
diminishing the quantity of light transmitted. The loss by 
absorption, in case total reflection prisms are used, will be 
correspondingly small, because the prisms themselves are small 
in comparison with the prisms of the spectroscope train. 

As regards the second objection, there is hardly need to 
consider it at all when such surfaces as Mr. Brashear, Mr. 
Clark or the Henry Bros, turn out are at our disposal. It is 
only necessary to refer to the magnificent definition of the 
Coud6 Equatorials and to remember that the diflficulties of 
making and supporting such large surfaces as are there used are 
many times greater than would ever be encountered in the instru- 
ments at present under discussion. It is to be noted, moreover, 
that all of the surfaces are behind the slit and receive only the 
radiations which pass through the latter, the energy in which is 
far too small to heat them appreciably, even in the case of solar 
work. 



242 F. L. O. WADSWORTN 

The third objection is of comparatively little importance in 
the case of visual observations, but of verj' great importance in 
the case of photographic or automatic bolographic observations, 
because in these we rely absolutely on the adjustment of the 
spectroscope remaining the same during the whole time of 
exposure, which in some cases may be several hours. The effect 
of any displacement of the different parts with respect to each 
other is greater in a spectroscope in which reflectors are used 
than in one in which no reflecting surfaces are involved, and under 
unusually severe conditions of use the former instruments would 
undoubtedly be less satisfactory than those of the usual form. 
But I think that the results which have been obtained by Young, 
Lockyer, Huggins, Kayser and Runge, Rubens and others with 
various forms of Littrow spectroscope in which from three to 
sometimes as high as ten or more reflecting surfaces were used, 
as well as my own experience with some of the forms which have 
just been described, quite justifies me in the statement that when 
these instruments are carefully designed and properly used no 
difficulty need be apprehended from a change of adjustment 
during the longest exposures. I have already spoken, in the 
earlier part of this article, of the admirable steadiness and con- 
stancy of adjustment of the large fixed-arm spectroscope of the 
Astrophysical Observatory at Washington ; an instrument in 
which the two reflecting surfaces, i. r., the prism-mirror and the 
concave reflector which serves as objective of the view telescope, 
are nearly four meters apart ; and, as I have stated in a previous 
paper, I have taken solar spectrum photographs (in the red and 
infra-red region) of between one and two hours' exposure, with 
the form of instrument shown in Fig. 4, in which the lines were 
perfectly sharp and showed no tendency to the broadening which 
would have resulted from any change in adjustment during that 
time. It is true that this time is short as compared with the 
exposures sometimes given on star spectra, but it may be 
remembered on the other hand that the conditions of use were 
unfavorable to steadiness in both the above cases, particularly the 
latter, in which the frame of the instrument was made entirely of 



THE MODERN SPECTROSCOPE 243 

wood. At night the conditions both as regards uniformity of 
temperature, and freedom from vibration and earth tremors, are 
very much better than during the day. I was, myself, formerly 
quite as skeptical as the majority of astronomers seem to be in 
regard to the use of reflecting surfaces in a spectroscope train, 
but, in view of the preceding experiences, I have not hesitated to 
employ them freely in my recent designs, whenever any decided- 
advantages could be attained thereby. 

I believe that the advantages of the use of concave mirrors 
in place of lenses especially have never been fully appreciated, 
or rather, perhaps, their disadvantages have been greatly over- 
rated. 

The form of spectroscope last described would seem to be 
particularly suitable for prominence spectroscopes and spectro- 
heliographs, one form of which is shown in Fig. 6. In this the slit, 
s, and the first reflecting prism, a, which is placed just behind it, 
are mounted together on a carriage, c, which slides on ways 
parallel to the direction of the reflected ray a a^. This makes it 
possible to examine any part of the Sun's surface or limb without 
moving either the solar image or the spectroscope. In order to 
automatically maintain the slit at the principal focus of the col- 
limating lens, the latter may also be mounted on a sliding 
carriage, connected to the carriage c by means of a steel cord or 
suitable system of levers. For the dispersion system we may 
use either a train of fixed prisms adjusted once for all to bring 
the C or K line into the field of the observing telescope, as 
shown in the figure, or, better, a train of adjustable prisms with 
independently movable reflector, like that used by Professor Hale 
in his last form of spectroheliograph.' Or, since the second 
reflector a^ is fixed in position and may be set so as to reflect the 
light from a in any direction, the axis of the collimating telescope 
may be placed at such an angle with that of the observing 
telescope that a grating or combined grating and prismatic train 
like that of Figs. 2, 2a, or 3 may be used. When used as a 
spectroheliograph the carriage, c, is also made to carry a photo- 
* A. and A, October, 1894; see also A. and A, March, 1893. 



244 



F. L, a IVADS WORTH 



graphic plate, which moves just behind the slit, s^ of the image 
telescope- The slit and photographic plate are, therefore, fixed 
rigidly with respect to one another, and to make the exposure it 
is only necessary to move the carriage, c c, which carries both, 
across the solar image which is formed on the first slit plate, 5, 
the image itself and all other parts of the spectroscope remaining 
•fixed in position. 

All lever systems and movable connections between the slit 
and photographic plate, or between the two slits, are thus entirely 
avoided, and as both slits are fixed in the axes of their respective 
telescopes there is no distortion of the photographic image 

C 




Fig. 7 



(except that due to the curvature of the line on the second slit). 
In fkct the result optically considered is the same as that pro- 
duced in the most recent spectroheliographs of Hale and Des- 
landres, in which the entire spectroscope is moved bodily across 
the solar image, and it was the description of these instruments 
[^Astronomy and Astro- Physics, November, 1894 ; VAstronotme^ June, 
1894)^ which suggested to me the preceding simple substitute. 

To avoid the necessity for moving the collimating lens with 
the slit carriage (although there is not really any serious 
mechanical or optical objection to this), the modification shown 
in Fig. 7 may be adopted. The ray from the first (slit) reflector, 
^, is received by a second double total reflection prism, ^, by which 

' The priority of suggettion of this form of instniment belongs to Professor Hale, 
who described it in an exhaustive article on the general subject of the spectrobelio- 
graph published in Astronomy and Astro- Physics for March, 1893, over a year before 
M. Deslandres' instrument was constructed. 



THE MODERN SPECTROSCOPE 245 

it is reversed in direction and returned to the second reflector, a\ 
fixed in the axis of the collimating telescope, B, which may be 
parallel to the axis of the observing or image-forming telescope, 
A^ or, as in the preceding cases, inclined to the latter at a small 
angle. 

The axis of the observing telescope. Ay may be immediately in 
line with the axis of the large telescope, and in this case the 
photographic plate is mounted immediately behind the slit and 
first reflector, making a very light and compact arrangement. 
The carriage, <:, which carries the slit, first reflector and plate is 
connected to the plate on which b is mounted by a simple arrange- 
ment of levers or multiplying gears, so that the latter moves with 
it in the same direction but at one-half the speed. This motion 
of the second reflector serves only to keep the distance s a b a' 
constant, and does not therefore need to be especially accurate 
either as to amount or parallelism, as any slight angular displace- 
ment of the double reflector, b^ is without effect on the direction 
of the reflected ray b a' . 

By removing the carriages c and b, moving out the slit j, 
until it occupies the position of s, and inserting an eyepiece at a' 
in place of the reflecting prism, the instrument is converted into 
an ordinary spectroscope. Conversely any spectroscope of the 
usual design may be simply and inexpensively converted into a 
spectroheliograph by the addition of the two carriages c and b 
and the system of reflectors a, b, a\ or by the addition of c and 
the reflector a' alone, if the former be directly connected to the 
collimating lens as in the preceding arrangement (Fig. 6). 

THE OBJECTIVE PRISM. 

The many advantages which an objective prism possesses 
over any form of compound sfar spectroscope as respects 
simplicity and high resolving power with a maximum brightness 
of spectra, would make its use more general, were it not for the 
difficulty of obtaining and mounting the very large prisms 
necessary for a telescope of even moderate size. 

The largest aperture which has been heretofore used with this 



246 F. Z. O. IVADSIVORTH 

type of instrument is that of the photographic telescope of the 
Harvard Observatory, although an objective prism has now been 
completed for the new Bruce photographic telescope which is. I 
believe, of the same aperture as the telescope itself, i. ^., twenty- 
four inches. 

Since the difficulty of obtaining homogeneous blocks of glass 
is very much less for thin than for thick plates, it would seem that 
in making a very large prism it might be advisable to build it up 
out of a number of comparatively thin prisms, as shown in Fig. 8, 
cemented together into one solid prism. Since there is no 
difficulty in obtaining thin plates as large as the disks which 




Fig. 8 

have been cast for the largest telescopes, it would seem perfectly 
practicable by this method to make a prism of an equally large 
aperture and of any refracting angle, and consequently of any 
resolving power that is desired. Another advantage of this 
method of construction would be that the central prisms might 
be made of glass of a greater density than the lateral ones and 
the advantages of the compound prism, as regards greater 
resolving power and dispersion for a given volume of glass and 
a given loss of light, realized. 

But in an objective prism we require usually only a mod- 
erate resolving power, the increase in aperture being advantageous 
simply on account of the larger quantity of light and greater 
brightness of the resulting spectra. Hence a better arrangement 
than the use of one very large prism would be the use of several 
smaller prisms of the required refracting angle (to produce the 
required linear dispersion at the focal plane of telescope) 
arranged ** in parallel " so as to cover the whole surface of the 
objective, as shown in Fig. 9. 

This enables us to utilize the entire aperture of even the 



THE MODERN SPECTROSCOPE 



247 



largest objective without an undue increase in the size and 
bulk of the prisms. In order that the optical performance 
of this system shall be perfect, it is necessary that the prisms 
shall be of the same material, that they shall have the same 
refracting angle, and that they shall each be set at approxi- 
mately the same angle to the axis of the telescope. Of these 
conditions the first and third are easily satisfied, and the degree 
to which the second may be fulfilled depends only on the skill 
of the optician. Since the two faces of plane glasses or sex- 
tant mirrors may be made parallel to within less than a second 
of arc, it would seem easily possible to work the refracting angle 





i" , ii.. 


^. 




, llM.. 

Ill- 


V 

Ii- '' 

|i' |,,,... 


\ lb... 


%■■■ i 


/•■ * 


"^^ 


ii. 


^ y ' 




Fig. 9 

of a prism to within this same limit of accuracy. This would 
mean that the angular deviation of the superposed spectra from 
the different prisms would differ from one another by less than 
half a second of arc, a quantity which is considerably less than 
the usual width of the image of the source as broadened by 
diffraction and aberration. No great degree of care need be 
exercised in setting the prisms at the same angle to the axis of 
the telescope, for a difference of 5 ' from the position of minimum 
deviation in a white flint prism of 30° changes the angle of devia- 
tion of the refracted ray by only about 0.3'. For this same reason 
changes in the relative position of the prisms during use, due to 
changes of temperature or other causes, will have little or no 
efiEect on the purity of the resultant spectral image. 

University of Chicago, 
January, 1895. 



Minor Contributions and Notes. 



THE DESIGN OF ASTRONOMICAL SPECTROSCOPES. 

In the first number of The Astrophysical Journal Professor 
Wadsworth, employing the methods of physical optics, and assuming 
constant resolving power as a basis of comparison, has discussed in a 
very complete manner the conditions on which the efficiency of spec- 
troscopes for astronomical purposes depends. In 1891 1 published in 
the Sidereal Messenger an elementary paper on the same subject, from 
a different standpoint, employing the methods of geometrical optics, 
and adopting constant dispersion as a basis of comparison. The pur- 
pose of the present note is to compare some of the conclusions which 
were reached in these two papers. 

The condition of constant resolving power is much the more philo- 
sophical of the two bases of comparison, but it is obviously unsuitable 
for the purposes of an elementary discussion. The observer generally 
has to deal with a given dispersion, and constancy of resolving power 
involves the variation of other quantities which are more naturally 
regarded as constant. The approximate methods of geometrical 
optics are also better suited to such a discussion, and they are sufficiently 
accurate for all cases which occur in astronomical spectroscopy except 
the case of the Sun, and possibly that of the stars.' For a thorough 
discussion the resolving power is more suitable as a basis, and too much 
value cannot be attached to the exact methods of physical optics, as a 
means of indicating the limit to which our results tend under the best 
conditions of observation. Whichever method of treatment is adopted, 
the conclusions, except in certain special cases, should be the same, and 
in general the conclusions of Professor Wadsworth and myself either 
agree, or difiEer on some point of comparatively small importance. In one 
respect, however, they differ very widely, and that in a matter of such 
importance that its further discussion seems to be very desirable. 
Whereas I have advocated the use of a small number of large prisms 
with correspondingly large spectroscope aperture (the limiting size to 

' In this case the width of the slit must be considerably greater than its theoretical 
value to obtain full illumination, doubtless on account of the motion of the image. 

248 



MINOR CONTRIBUTIONS AND NOTES 249 

be determined by practical considerations of cost and weight), Pro- 
fessor Wadsworth concludes that a spectroscope of proper construction 
should have a small aperture and a large number of small prisms. He 
also concludes that correct lines have not been followed in the con- 
struction of existing telescopes, or rather that more efficient instru- 
ments could be built on different lines, provided the plan of construc- 
tion should include the telescope itself. 

As stellar spectroscopy is becoming a part of the regular work of a 
well-equipped observatory, it is a matter of importance to settle this 
question as to the lines which should be followed in the construction 
of instruments. Now I believe that the best spectroscopes of modern 
construction are more correctly designed than Professor Wadsworth 
seems to think, and that they could not advantageously be replaced by 
instruments of the construction which he recommends. It is of course 
true, as Professor Wadsworth says, that spectroscopes have always been 
made to fit previously existing telescopes, but I doubt whether a radical 
change in any part of the design would be advantageous. 

My reasons for this opinion, which are all of a practical character, 
are given below ; at first I will leave the case of a star out of consider- 
ation. Since the brightness of the spectrum mth constant resolution is 
independent of the aperture of the spectroscope, it is evident that a 
given resolution can be obtained by employing either a small number 
of large prisms or a large number of small prisms. Which method 
should be employed becomes after all principally a question of opinion 
based on practical experience. The advantages of the latter method 
are, according to Professor Wadsworth, greater lightness, stability, and 
cheapness of construction, with some others of less importance. The 
greater lightness may be conceded, although a spectroscope of from 
I to i^ inches aperture is by no means a heavy burden for a 12-inch 
equatorial, and it can be attached or removed without difficulty. Under 
the condition of similar form, the advantage with respect to rigidity 
might lie with the smaller instrument, but with a large number of small 
prisms, each of which would be subject to slight displacements during 
a long exposure (for prisms cannot be very tightly clamped) the small 
instrument would probably be less rigid than the large one ; moreover, 
there is no difficulty in making a spectroscope of the size mentioned 
above so rigid that displacements due to flexure are of less importance 
than those due to changes of temperature under the ordinary circum- 
stances of observation. The cost of construction can hardly be regarded 



250 MINOR CONTRIBUTIONS AND NOTES 

as following even roughly any particular law with respect to size, and the 
larger number of small pieces required in Professor Wads worth's con- 
struction, with their accurate fitting, might easily raise the cost to that 
of the larger form. 

With a large number of prisms, errors in the surfaces are not likely 
to be compensatory, as in ordinary optical work the tendency is invari- 
ably to make the surfaces convex.' In the present state of the optician's 
art, however, errors in the surfaces are less to be feared than irregular- 
ities of density in the glass. The effect of such irregularities would, 
perhaps, be less with the small prisms. Finally the loss of light by 
reflection would be somewhat less in the case of the large prisms.' 
So far, then, a departure from the usual construction of a spectroscope 
does not seem to be advisable. 

But the most serious objection to the small aperture and high dis- 
persion is, perhaps, the following: Since the linear extent of the 
spectrum cannot exceed a certain limit on account of the faintness of 
the light, it would be necessary to use a camera objective of very short 
focus. As compared with the other form of spectroscope having equal 
resolving power, the field would therefore be very small, and only a 
short range of spectrum could be sharply photographed on a single 
plate. Except for a very few special purposes this would be decidedly 
objectionable. I am at present engaged in some investigations in 
which the small field obtained even with a long focus camera is a seri- 
ous inconvenience. 

In the case of a star, which may now be considered, a telescope of 
the largest possible aperture \& of course desirable. Professor Wads- 
worth advocates the use of a reflector, which unquestionably has some 
great advantages, the greatest of all being the absence of chromatic 
aberration. To this reflector he would give a very short focal length, 
principally in order to contract the linear dimensions of the diffraction 
pattern, and thus allow the use of a correspondingly narrow slit. But 
since the angular aperture of the collimator must be equal to that of 
the telescope, and it is impracticable to make a sufficiently good lens 

' The limiling error of the surfaces of the prisms made by Brashear for the Alle* 
gheny spectroscope is only ^ \ but errors amounting to several wave-lengths are fre- 
quently met with in ordinary prisms. 

'About 25 per cent, less, according to Professor Wadsworth's tables, if we com- 
pare three prisms with six prisms of half their size. The advantage is less with a 
greater number of prisms. 



MINOR CONTRIBUTIONS AND NOTES 25 1 

of the large angular aperture required, the usual collimating lens must 
in that case be replaced by a parabolic mirror. This seems to me to be 
a highly objectionable arrangement, as according to all experience a 
mirror cannot be depended upon when stability is required ; in fact, a 
reflector is as much out of place in a spectroscope designed for long 
exposures as it would be on a meridian circle. If, however, we abandon 
the use of the reflecting collimator, we must again fall back upon the 
usual construction, with an angular aperture of something like 1:15. 

In making the above remarks I would not at all be understood as 
underestimating the valuable discussion of Professor Wadsworth, which 
contains many novel and interesting features; but I wish to point out 
that experience cautions us not to accept the conclusions too hastily, 
although they are founded on theoretical principles which are undoubt- 
edly correct. Perhaps it would be advantageous to make a compromise 
between conflicting requirements by using a reflecting telescope of 
unusually large angular aperture, say i : 10 or i : 8, and a refracting col- 
limator with triple objective. The three lenses would allow the focal 
length to .be made short with sufficient exactness in the corrections, 
and the construction of the lens would be facilitated by the fact that 
the question of fleld does not have to be considered. 

An interesting point raised in Professor Wadsworth's paper may 
appropriately receive notice here, although it is not immediately con- 
nected with the present subject. Professor Wadsworth deduces the 
width of the nebular lines from certain photographic observations of 
mine on the Orion nebula, although with due reservation as to the 
reliability of the data. The data are reliable in the sense that they 
represent correctly the observed facts, but the method of observation 
is altogether too rough to give results of value. Briefly reviewed, the 
entire method is as follows : The brightness of a spectral line increases 
up to a certain limit as the slit of a spectroscope is widened, and then 
remains constant. The limiting slit-width is, in the case of monochro- 
matic light, ^o = T • I^ the light includes wave-lengths ranging through 

AX, Professor Wadsworth has shown that the limit is sj =- (X + tAX). 

Hence if the limiting slit-width can be determined experimentally 
the value AX can be found. In photographing the spectrum of 
the Orion nebula with the Allegheny spectroscope, I found that the 
denstty of the lines on the negative fell off sensibly if the slit was made 



252 MINOR CONTRIBUTIONS AND NOTES 

narrower than .001 inch, when it was still three times the theoretical 
limiting width for full illumination by monochromatic light. From 
these and other data relating to the instrument, Professor Wadsworth 
obtained the value AX=2.6 tenth-meters. 

Now as a matter of fact the width of the nebular lines is much less than 
this. In my observations of nebulae at the Lick Observatory the width 
of the lines as ordinarily observed was only 0.35 tenth-meter, and I 
have seen the lines when their width was much less than this, and prob- 
ably not more than o.i tenth- meter. Probably they are as fine as the 
lines of hydrogen at the lowest possible pressure. The explanation of 
the discrepancy is, I think, that which I gave at the time ; it depends 
upon a peculiarity of the photographic action which, as shown by other 
experiments, may produce very considerable effects. The structure of 
a photographic plate seems to be too coarse for such experiments as 
these, or at least for such applications. It is to be hoped that further 
observations on this interesting subject will be made with more suitable 
apparatus. j^^^s E. Keeler. 

NOTES ON SILVERING SOLUTIONS AND SILVERING.' 

Since the first introduction, a half century ago, of Liebig's method 
of silvering glass by deposition from the solution of a silver salt, a 
number of modifications of the process (consisting usually in some 
variation in the reducing agent employed to precipitate the silver) have 
been proposed, some designed to cheapen the cost of the method by 
securing a larger percentage of deposited silver, and others to render 
the deposited film harder and more enduring. Of this latter class that 
which was originated by Brashear is one of the most successful, as it 
gives a film so hard and adherent that it may be rubbed vigorously 
with the hand or with a pad of cotton, while it is still wet from the sil- 
vering bath, without injury. A description of the process was pub- 
lished some years ago by Mr. Brashear' and is now very generally used 
by professional makers both in the United States and in England, but 
it is not as generally well known among scientific men as its merits 
deserve. It was described to me about a year ago by Mr. Brashear 
himself, and since that time it has been used exclusively in silvering 

'From the Zeiischrifi fiir Instrumentenkunde, January, 1895. 

* English Mechanic^ 1883; also in a pamphlet, "Silvered Glass Reflecting Tele- 
scopes and Specula," by J. A. Brashear ; Best & Co., Pittsbni^g, Pa. 



MINOR CONTRIBUTIONS AND NOTES 



253 



the large mirrors (one of which is 36^" in diameter) used on the 
siderostat and spectro-bolometer of the Smithsonian Astrophysical 
Observatory. It has given excellent results, and for the benefit of those 
who are unacquainted with it I will describe it briefly, giving the pro- 
portions both in grams and cubic centimeters and in the English 
units used by Mr. Brashear. 

The composition of the reducing solution is as follows : 
Loaf sugar or rock candy - 90 gms or 840 grains 
Strong nitric acid (sp. gr.= i.22) 4 cc or 40 grains 

Alcohol 175 cc or 35^ ounces 

Distilled water - - - 1000 cc or 25 ounces 
This is made up by dissolving the sugar in the distilled water and 
then adding the alcohol and the nitric acid. It should be prepared at 
least a week before it is used and, unlike most solutions for this pur- 
pose, the longer it stands the better it gets. A quantity sufficient to 
last a year or more may therefore be made up at one time. 

The silver solution is an ammoniacal solution of the oxide, precip- 
itated as in the ordinary process from the nitrate, to which just before 
using is added a solution of caustic potash in the proportion of % 
gram of potash (KOH purified by alcohol) to i gram of the silver salt. 
If very much, silvering is to be done, both the ammoniacal solution of 
silver and the caustic potash solution may be kept in stock and mixed 
as required, but better results will be obtained if this part of the silver- 
ing bath is made up as it is wanted. 

The amount of silver nitrate, potash and ammonia, and the corre- 
sponding quantity of the reducing solution required for different sizes 
of mirrors will be about as follows : 



For Minora 


Ai«a 


Silver Nitrate 
(Ag, NO.) 


Caustic Pbtash 

(KOH) 


Ammonia > 
NH40H + H,0 


Reducing 
Solution 


30 cm diam. 
25 " " 
20 " 
15 " 
10 " 
5 " " 


707 sq. cm 
491 
314 
177 

78.5 

19.6 


II 
7 
4 
1.8 

0.5 


7.5 gPas. 

5.5 

3.5 

2.0 

0.9 

0.25 


12 cc 
9 
6 

\^ 


85 CC 

65 
40 

10 

3 



In English units these quantities correspond to 120 grains of nitrate 
of silver and 60 grains of potash for a mirror %% inches in diameter. 

' The amount of ammonia will of course vary with the strength of the solution. 
The quantity here indicated is for ammonia of sp. gr. about 0.88. 



254 MINOR CONTRIBUTIONS AND NOTES 

The bath is made up as follows : The silver nitrate and the potash 
are dissolved separately, each in about loo*^*" of water per gram of 
salt. To the silver solution is added about one-half the ammonia, and 
the remainder is diluted with distilled water in the proportion of i to 
5, and then added more slowly until the silver precipitate is barely 
redissolved. During the last part of the process the solution should 
be constantly agitated, and the vessel which holds it should be occasion- 
ally lifted or shaken so as to wash down the sides. (A Florence flask is 
much to be preferred to a beaker for this operation, because of the 
greater facility with which it may be handled.) The potash solution is 
now added and mixed thoroughly, and if a precipitate remains, dilute 
ammonia is added until it is not quite redissolved, using the same 
precautions as before. At the end the liquid should have a slight 
brownish color, indicating the presence of a little free silver oxide. It 
should be allowed to stand for a few minutes, and then if there are 
many floating particles it is filtered through coarse filter paper or 
cotton, after which it is ready for use. 

A slightly different mode of procedure is recommended by Mr. 
Brashear. About ^^ of the original silver solution is reserved, and, 
after the balance has been treated as already described, the reserve 
silver is added slowly until another distinct precipitate is formed ; then 
a little more ammonia until it is redissolved, then a few more drops of 
silver, and so on until all the reserve silver has been added ; taking 
care to make the last addition silver solution and not ammonia. 

If care is taken to use as dilute a solution of ammonia as recom- 
mended above, and to agitate the solution thoroughly during its 
addition, the first method will be found perfectly satisfactory. It is, 
however, well for beginners to adopt the latter plan until they learn to 
recognize from the appearance of the solution the presence of free 
silver oxide. 

** // is useless to attempt to silver without having a slight excess of silver 
in the solution. ^^ 

The solution having been filtered as above described (if there are 
no floating particles, this filtering will be unnecessary) the required 
amount of reducing solution is added, the whole thoroughly mixed and 
poured into the dish in which the silvering is to be done, and the 
mirror, which has previously been cleaned (of which more hereafter), 
immediately immersed face up or down, as the operator prefers. I 
myself always prefer to silver face up, as the progress of the deposit 



MINOR CONTRIBUTIONS AND NOTES 255 

may then be watched, and arrested when it has proceeded far enough. 
When silvering* face up, however, the solution must be kept in constant 
motion to prevent particles of precipitated silver from settling on the 
glass surface. 

In a few minutes after mixing the bath turns a dark brown color, 
which, as the operator proceeds, gradually becomes lighter and lighter, 
until at length it is nearly clear again. At a temperature of 70° F., 
which is best for securing good results (if much lower than this the 
film will be too thin, and if much higher, too soft), the operation will 
be finished in from ten to fifteen minutes. The mirror is then lifted 
out, placed in an inclined position under a stream of clean water, and 
the whole surface rubbed vigorously with a pad of clean absorbent cotton 




Fig. X 

until the white film on the surface of the silver is entirely removed, and 
the whole surface is bright and clear. The mirror is finally set on edge 
on a sheet of blotting paper in a warm dustless place and allowed to 
dry. If the operation has been successful, a bright hard surface will 
be obtained which will need no polishing, and will therefore be free 
from the minute scratches always produced by the polishing pad, no 
matter how carefully the latter may be prepared and kept. One 
essential condition to success is the use of clean wash water^ not neces- 
sarily distilled water, but water which is at least free from free alkalies 
and acids, and from any suspended sediment. 

The foregoing process is of especial value where non-diffusive coats 
of silver are desired, for example, on concave mirrors of that form of 
Littrow spectroscope recently described by the author.' 

« " An Improved Form of Littrow Spectroscope," F. L. O. Wadsworth, Phil. Mag, 
July,* 1894* 



256 MINOR CONTRIBUTIONS AND NOTES 

When half silvering, 1. ^., when a very thin semi-transparent coat — 
such as is used on telescope objectives for solar observations, or on the 
"separating glass" of the "interferential refractometer " — is required^ 
the old Rochelle salts process will give the most satisfactory results as 
regards uniformity, although the film is less enduring than that pro- 
duced by the Brashear process. As the directions given in the books 
in regard to this process are generally meager and often misleading, 
the following notes on certain precautions necessary in order to insure 
success in this, the most difficult of all silvering operations, may be of 
service to those who have had but little experience in the art. 

In the first place pure chemicals, while not absolutely essential, are yet 
of considerable advantage, and it is best to purchase them C. P. of some 




Fig. 2 

reliable dealer in chemical supplies. If the ordinary silver nitrate and 
Rochelle salts of commerce are used, recrystallization is desirable. In 
making up the silver nitrate solution great care should be taken to 
avoid an excess of ammonia, by leaving the solution decidedly brown 
before filtering. In making up the reducing (Rochelle salts) solution I 
first bring the distilled water to boiling, and then add first the silver and 
then the Rochelle salts, both of which have previously been dissolved 
in the smallest possible quantity of boiling water. The boiling is then 
continued from twenty minutes to half an hour, or until the gray pre- 
cipitate has collected together in the form of a compact powder at the 
bottom of the flask, leaving the supernatent liquid nearly clear. It 
should then filter perfectly clear, and remain so after cooling. Much 
seems to depend on the length of time of this boiling ; as regards the 
performance of the silvering bath, in general the longer the boiling 
the more rapidly will the deposit take place, and the more uniform it 
will be. 

Cleaning the glass for silvering, — It may truly be said that no other 
one condition of success is one-half as important as a proper cleaning 
of the surface to be silvered. In four cases out of five a failure to 
secure good results is due to improper cleaning. The necessity for this 



MINOR CONTRIBUTIONS AND NOTES 2$? 

is especially apparent in half silvering, for the deposit must here be 
absolutely uniform from the very beginning of the operation. Merely 
bathing the surfaces with acid, potash, and alcohol in succession, as 
recommended in the books, is far from sufficient, unless the surface be 
unworked.' The best plan is to wash the surface thoroughly with a 
pad of absorbent cotton. Then rinse in clear water, and transfer to a 
dish filled with strong nitric acid. Again go over the whole surface 
thoroughly with a pad of cotton, held on the end of a bent glass rod of 
the form shown in Fig. i. The surface must be rubbed hard, not 
merely lightly brushed over. If care is taken to flatten down and 
round over the end of the rod, as shown in the figure, and to pick out 
a piece of cotton free from any gritty particles, there will be no danger 




Fig. 3 

of injuring the surface. Then pour off the acid or transfer the glass to 
another dish filled with a strong potash solution, and repeat the rub- 
bing. Finally rinse, and place in a dish of pure distilled water until 
ready for silvering. 

The use of the alcohol as recommended by the books is not only 
unnecessary but, unless the glass be very thoroughly washed sub- 
sequently, is actually detrimental. If any considerable amount of 
organic matter is present, it should be removed ^^<;r^ commencing the 
cleaning, either by washing with alcohol, or better by a bath of sulphu> 
ric acid to which some permanganate of potash has been added. For 
cleaning large mirrors Mr. Brashear recommends that, after being 
treated with nitric acid and potash, the surface be rubbed with prepared 
chalk until it is thoroughly clean and dry. It is then either washed 

' It is a remarkable fact, which has probably been noticed by all who have had much 
silvering to do, that it is much more difficult to silver on a worked (i.r., ground, 
polished) surface than on one from which the natural blown surface has not been 
removed* 



258 



MINOR CONTRIBUTIONS AND NOTES 



again with water, or left dry as it comes from the chalk polishing until 
the silvering bath is ready. 

When the surface is properly cleaned, the distilled water will wet 
and flow over the whole surface uniformly ; if it is not, it will collect 
in drops on the plate. If it does this, the cleaning operations must be 
repeated. 

It is essential that the utmost cleanliness be observed in all of these 
operations. They should be conducted preferably in glass or porcelain 
vessels, and the fingers should never be allowed to touch the surface 
which is to be silvered. If small, the plates are handled by means of 




Fig. 4 

glass tongs or. supports like those shown in Fig. 2. If the plate 
is large it is necessary to use two of the latter, one on each side, or 
better to make a glass frame with two handles, similar to that shown in 
Fig. 3. For still larger mirrors, a glass dish with a stopcock in the 
bottom, from which the different washing fluids may be drawn in suc- 
cession, may be used. The mirror is supported in this case by means 
of small glass rods which will permit of perfect washing. A shallow, 
stoppered bell glass is perhaps the most available commercial article. 
This is supported on two blocks of wood so as to lift the central stop- 
per, into which the discharge tube is cemented by means of parafiine, 
away from its support. For the sake of economy the glass dish should of 
course be but very little larger than the mirror. (Fig. 4.) In case 
mirrors of irregular shapes are to be silvered, special dishes of just the 
right size may be made by roughly nailing together a box of wood of 
the required size and dimensions, and then thoroughly coating the 
inside with very hot melted parafline. If the mirror be very large 



MINOR CONTRIBUTIONS AND NOTES 259 

indeed, the most convenient as well as the most economical way is to 
let the surface to be silvered itself be the bottom of the silvering dish, 
and form the sides by passing around the edges of the mirror a strip 
of paraffined paper, which is held in place by a rubber band or cord. 
The paper is cemented to the edge of the mirror by rapidly passing a 
hot iron around the latter. A shallow dish is thus formed, which serves 
to hold the cleaning and finally the silvering fluid. Or by using care 
the band may be put in place after cleaning, and used only for the silver 
bath, since it is only this that needs to be economized. 

If the mirror is to be silvered face downward — a method preferred 




Fig. 5 

by some because less care is necessary during the silvering — some 
method of support must be adopted which will lift the face about a 
centimeter above the bottom of the silvering dish. In this case the 
method of support which I prefer, as more cleanly and less wasteful 
of space than any other, is the use of three wedge-shaped blocks of 
paraffined wood placed in the bottom of the silvering dish. When 
finished the mirror is removed by placing the hand covered with a 
sheet of blotting paper on the back, and then inverting the dish. If 
the mirror is too heavy to be held in one hand, both may be used while 
an assistant inverts the dish. Another method which I have never 
used, but which perhaps might be equally cleanly and efficient, would 
be the use of a suction clamp applied to the back of the mirror. 

(Fig; 5) 

To briefly recapitulate, the points essential to success in silvering 
are : (i) a thorough and systematic cleaning of the surface ; (2) reason- 
ably pure chemicals, and a silver-nitrate solution containing an excess 



26o MINOR CONTRIBUTIONS AND NOTES 

of silver; (3) uniformity of temperature between the silvering bath and 
the mirror, preferably from 15° to 20° C. ; (4) the use of clean water, 
and plenty of it, in all the stages of operation, especially in the final 
washing. 

The methods of operation described in this paper are probably 
most Of them already familiar to those who have had experience in this 
work. It is not for them that this paper is written, but for those who, 
because of inexperience, have had difficulty in securing uniformly good 
results. If it is of any service in assisting them to a knowledge of a 
better method (and I certainly consider Mr. Brashear's method far 
superior for general use to the old Rochelle salts process), or of more 
convenient ways of working, the object of the author will have been 
accomplished. 

F. L. O. Wadsworth. 



ON BRESTER'S VIEWS AS TO THE TRANQUILLITY OF 
THE SOLAR ATMOSPHERE. 

In the December (1894) number of Astronomy and Astro- Physics 
(p. 849) Mr. Brester has replied to certain criticisms of his Theory of 
the Sun which I had advanced in the number of the preceding August. 
A final word from me in reply to these remarks may, perhaps, be 
permitted. 

To completely justify my objections it is only necessary, I think, to 
quote the words used by Mr. Brester (" A short Review of my Theory o^ 
the Sun," Astronomy and Astro-Physics, 13, 218). 

" The interior tranquillity of the Sun is not to be thought of merely 
as the conditio sine qua non of its permanent stratification, but as a 
priori much more probable than the ordinary theory of violent agita- 
tion A quiet interior should, therefore, not surprise us. 

. . . . If, in general, solar phenomena teach us that the Sun is in 
a state of repose, and that this quiet is such that the Sun, in spite of its 
gaseous state, presents the appearance of a solid. . . ." 

When one reflects that from purely physical considerations vertical 
movements in a stable atmosphere, either of a heavenly body or of our 
Earth, are not comparable with horizontal movements, one cannot 
speak of " a state of repose " of the Sun, if, as I have shown in the 
article referred to, storms of more than 100 meters a second accompany 



MINOR CONTRIBUTIONS AND NOTES 26 1 

the most ordinary phenomena on the Sun's surface — storms which 
exceed in violence by threefold our hurricanes. My objections seem, 
therefore, justified.' 

Since Mr. Brester very naturally desires to dispose of my theory of 
the constitution of the Sun's atmosphere and the cause of the Sun-spots, 
I feel it necessary to reply briefly to his objections. 

Mr. Brester seeks especially to disprove the existence of a very 
small density at the surface of the photosphere, and on the following 
grounds : 

In the tirst place the radiation from different parts of the Sun's 
surface shows, according to Vogel's measures, a considerable increase 
of the general absorption towards the limb, an absorption which varies 
with the color. Seeliger has shown that this necessitates the assump- 
tion of a medium of high refractive power. I have myself alluded to 
this apparent inconsistency in my paper, "Ueber die Ursache der 
Sonnenflecken " {SUzber. der Wien, Ak. d, Wissensch. p. 17, 1893). It 
finds, however, a ready explanation in the condition of the photosphere. 
A rare atmosphere in which flying particles are suspended fulfils the 
conditions of a highly refracting medium in the most complete 
manner. 

The second objection brought forward by Mr. Brester, namely, 
that the great height of the Sun's atmosphere presupposes also a great 
density at the surface, no longer applies when one remembers that the 
atmosphere of a rotating heavenly body only reaches a limit at the point 
where centrifugal force equals that of gravity. A great extension of 
the atmosphere, even such as that of the Sun's atmosphere, denotes a 
high temperature at the surface and nothing more. 

Mr. Brester refers also to the validity of Kirchhoff's law, which I 
have used in considering the glowing gas of the Sun's atmosphere, 
and cites in this connection certain researches on glowing gas ; while 
Paschen somewhat later has been led to quite different results, proving 
the dependence of the light emission on temperature. The question 
is in any case an open one. But these researches have little signifi- 
cance for the glowing gas on the Sun, since under the high tempera- 
ture which exists there all chemical action must be excluded. Kirch- 
hoff's law is, as he himself has shown in his well-known memoir, a 
necessary deduction if the quality and intensity of the light rays 

' I regret that in my paper, above referred to, I have by mistake interchanged the 
words "sphencal " and "flattened" (p. 583). 



262 MINOR CONTRIBUTIONS AND NOTES 

depend only on the nature and temperature of the radiating body. 
The application of Kirchhoff's law seems, therefore, at all events 
admissible. ^^^^ ^^^ Oppolzer. 

Munich, January, 1895. 



The Variable Star J4/6 S Velorum. — Professor Roberts explains his 
observations of this star {A, J,, No. 327) by supposing that the variable 
is a dark central star, with a brighter, but much smaller companion, 
revolving around it. The magnitude of the primary is 9".25, of the 
companion about 8 ".05. When the star is at a minimum, we have the 
light of the primary only, and the variable remains for over six hours 
at 9'".25. At a maximum we have the combined light of both stars, that 
is, 7 ".85. It is, he says, much to be regretted that the star cannot be 
passed under the scrutiny of the spectroscope. Such an examination 
would yield a most valuable train of facts, and would raise probable 
results into actual knowledge. It is to be hoped that ere long we will 
have in the southern hemisphere a spectroscope so adapted, and so 
powerful, that in half a dozen nights we will have settled, without the 
prospect of a doubt, the motions of this peculiar binary system ; and 
above all, the motions, relative masses, distances, and hence parallax 
of a, a, Centauri. 

As a 24-inch photographic telescope, to be provided with both an 
object-glass prism and a slit spectroscope, has been offered by Mr. 
McClean to the Cape Observatory, and will presumably be accepted by 
the Government, the time when a powerful stellar spectroscope will be 
permanently established in the southern hemisphere is not far distant. 
Unfortunately, even such an instrument as this could hardly deal with 
the motion in the line of sight of an 8"'-9" star. 

J. E. K. 



Reviews. 

Atude stir U spectre de tetoile variable S Cephei. A. B6lopolsky. 
Bulletin de FAcadUmie . . . de St. Petersbourg, \ig^, Novem- 
ber, No. 3, pp. 267-306. 

In A, N, z^, 281 B^lopolsky announced that this star is subject 
to an orbital revolution, and stated the provisional results of the spectro- 
graphic observations of its motion in the line of sight. A translation 
of this article was published in the February number of this Journal 
(pp. 160, 161). The present memoir contains the observations thus far 
obtained, with a discussion of the resulting orbit. 

For the very important purpose of investigating the spectra of 
variables which reach the sixth magnitude at minimum, a new spectro- 
graph, with a compound prism, was constructed and attached to the 
30-inch Pulkowa refractor. Under the best atmospheric conditions an 
exposure of an hour was sufficient to bring out the spectra of stars 
nearly as faint as the sixth magnitude. The portion of spectrum 
included was usually from about X4500 to X4300, the prism being 
adjusted for the rays at X 44 10. 

The comparison spectra employed were those of iron and hydrogen, 
the iron lines being impressed in about thirty seconds at the middle 
of the exposure, while the hydrogen light was thrown upon the slit at 
the beginning or end of the exposure. Unfortunately the Hy line 
seemed, in some instances, to be displaced with reference to the iron 
lines, thus giving different results for the two comparison spectra; we 
shall refer again to this discrepancy. 

Two methods of measuring the spectra were employed, — Vogel's 
first method, that of measuring relatively to a solar plate, and direct 
comparison with the iron lines at X4405 and A4415. The lines visible 
on each plate are enumerated in detail, with a description of their 
appearance. Two plates were obtained on most evenings of observa- 
tion, and each was measured by the two methods. 

The spectrum of 8 Cephei is of Type Ila, but some lines which are 
narrow and faint in the Sun appear broad and strong in the star. No 
marked changes in the character of the spectrum were noted as 

263 



264 



REVIEWS 



dependent upon the variation of the star's light, except, of course, the 
increased intensity of the spectrum at a time of maximum. 

The star is a variable of short period (5** S^'.S), ranging from 3.7 mag. 
to 4.9 mag., but the maximum occurs i'*i4*'.6, or only one-third of the 
period, after the minimum. 

The velocities in the line of sight, based upon the displacement in 
respect to the iron lines, and reduced to the Sun, are given below, after 
reduction from units and tenths of the lieue geographique to the nearest 
half kilometer (i /. g, = 5^.56): 



Panama M. 


T. 


Vek>city 


Pnlkowa M.T. 


Velocity 


d 


h 


km 


d 


h 


km 


1894 Aug. 3 


II. 


-24.5 


1894 Aug. 17 


10.8 


+ 0.5 


4 


10.8 


-17.0 


24 


9.6 


-28.0 


5 


II. 


-22.5 


25 


9.9 


-18.5 


6 


10.8 


-1- 4.0 


Sept. I 


9.5 


- 3-5 


8 


10.6 


— 22.0 


3 


9.4 


-27.0 


9 


10. 


— 16.0 


5 


9.3 


— 16.0 


12 


10.7 


~ II. 


6 


9.2 


- 6.5 


M 


10.4 


-18.5 


7 


8.9 


- 4.0 


16 


10.4 


— 10. 


II 


9.0 


- 9.5 



The mean error of the measures is - - - ±: 5^".5 
The mean error for each day - - - • i 3 
The mean error of the deduced velocities - d= 3 

It will be noted that the range in velocity is very much less than in 
the case of any of the spectroscopic binaries heretofore discovered, but 
it is quite too great to be due to errors of measurement. 

The orbit is computed by the valuable method first proposed by 
Rambaut {M, N, 51, 316-330), and more fully worked out by Lehmann- 
Filh^s {A, N, 136, 17-28). The period of revolution is taken to be the 
same as that of the star's variation. The velocity of translation of the 
system is found to be — 14*^ (approach). The other elements of the 
orbit are : 

tfi = 90^, the angle between the ascending node and the 
radius vector at the point where the sight-line 
component of the orbital velocity is o. 
« =88^, the angle between the ascending node and the 

perihelion (periastron). 
£ = 0.514, the eccentricity. 

Ig fL = 0.0678, the daily motion, in degrees = 67^.08. 



REVIEWS 265 

a sin i = 990,000^"* the projection of the semi-major axis. 

/ is not determinate, but in view of the probable eclipse it 
cannot be far from 90°. 
T =1.05 days after a minimum, is the time of periastron 
passage. 

The major axis of the orbit evidently lies nearly in the line of 
sight, periastron being on the farther side from the Earth. 

An ephemeris was computed from these elements, and it shows a 
fairly satisfactory accord with the observed velocities. 

A curve is given exhibiting the variation in the velocity with the 
time, and also the light-curve of the star after Oudemans* observations. 
The two agree quite well, but the interval of 1.05 days between the 
time of minimum and the passage of the periastron is not explained. 
The ephemeris of the star's variation, taken from the Annuaire du 
Bureau des Longitudes^ differs only very slightly from Chandler's ele- 
ments, and is probably quite accurate. 

In order to examine the relative displacement of the iron and 
hydrogen lines on certain of the plates, B^lopolsky has computed the 
velocities referred to the Hy line and compared the results with the 
ephemeris, with less satisfactory results. His conclusion is that the 
Hy line cannot in this case give as reliable determinations of veloc- 
ity as the iron lines because usually it was imprinted on but one 
of the two spectrograms taken on a night, because it was too far 
away from the portion of spectrum used for measurement and was not 
the line for which the prism was set at minimum, and because the 
exposure to the hydrogen spectrum took place either at the beginning 
or end of the exposure to the star, and not at the middle time as in 
case of the iron spectrum. The discrepancy is an unfortunate one and 
necessarily diminishes our confidence in the accuracy of the results. It 
is to be expected that in future work with this instrument its stability 
will be increased, and the method of obtaining the comparison spectra 
be so modified as to obviate this uncertainty. 

Of course the present observations form only a beginning of the 
researches which will be necessary to fully explain by orbital revolu- 
tions and eclipses the peculiar variations of the light of this star. The 
case is evidently much more complicated than that of the Algol vari- 
ables, although perhaps simpler than that of P Lyrae. 



266 REVIEWS 

The Luminosity of Gases. III. A. Smithells. Pkil. Mag,^ 
January, 1895. 
The second paper of this important series was noticed in A, and 

^' i3i 587-588, 1894. 

Of the present discussion the immediate subject is the various 
flame-spectra of copper oxides and chlorides, and of gold trichloride. 
One can here only hint at the method of study and at the results 
obtained. 

The Bunsen burner, fitted with a sprayer, a "saturator" and a 
cone-separator, furnishes a luminous flame, which is then examined 
with a two-prism spectroscope. The observed spectra are mapped in 
the following very happy manner, viz. : the ruled scale is one of equal 
parts, but each division, instead of denoting an equal increment of 

wave-length, X, denotes an equal increment of the value — • The 

A 

consequence is that the wave-length of each line is at once reducible 
to AngstrSm units, while the spectrum as a whole preserves its pris- 
matic appearance, the lines at the blue end being crowded together. 

As to results, the essentially new feature of this work is the discov- 
ery of the particular copper compound to which each of the various 
copper spectra are due. 

Following are the author's conclusions as stated by himself : 

" I. When cupric chloride is introduced into a flame, three sub- 
stances are formed : metallic copper, cuprous chloride, and cuprous 
oxide. The first of these gives a bright yellow flame and a continuous 
spectrum ; the second a bright blue tint and brilliant spectrum of bands 
and lines ; the last a green tint and a spectrum of not very strongly 
developed bands. Under certain circumstances cupric chloride may 
exist in a flame, when it gives a feebly ruddy tint and a continuous 
spectrum. 

"II. Gold chloride gives a flame-spectrum only in presence of an 
excess of chlorine or of hydrochloric acid and oxygen. 

" III. In the above cases the development of a spectrum is con- 
comitant with chemical changes affecting the substance concerned ; a 
fact in harmony with the view as to the origin of flame-spectra advo- 
cated by Pringsheim." 

The evidence which Professor Smithells adduces for each of the 
above statements is direct and clear ; but at the same time it is more 



REVIEWS 267 

than likely that many of his readers will refuse to admit that his evi- 
dence is conclusive. To illustrate, one is not quite certain that the 
products which condense on the outside of a vessel of cold water held 
in a flame are exactly the compounds which exist in the flame and give 
it luminosity. 

Concerning the subject of luminosity in general, and the Bunsen 
flame in particular, this paper only confirms a growing suspicion that 
one has here to deal with a vastly more complicated piece of mechanism 
than has hitherto generally been supposed ; and one is driven to wish 
that the problem might be attacked by the photographic method and 
large dispersion. H. C. 

Popular SciefUific Lectures. Ernst Mach, Professor of Physics in 
the University of Pragfue. Translated by Thomas T. 
McCoRMACK. (Open Court Publishing Co., Chicago, 1895.) 

The production of a popular scientific lecture of the highest order 
is to be reckoned among the later achievements of a profound scholar. 
The task requires a breadth of interest which the specialist has some- 
times lost early in his career ; it requires also a style which is lucid 
without the introduction of tedious detail, a style which is accurate with- 
out dependence upon formulae. These demands are thoroughly satis- 
fied in the volume before us. The lectures are not so strikingly orig- 
inal as some of Maxwell's nor so profound as those of Kelvin ; but 
they are truly remarkable in the insight they give into the relationship 
of the various fields cultivated under the name of Physics, and the 
book cannot fail to hold the attention of every reader who is inter- 
ested in the history of the development of Physics. 

The twelve lectures chosen for translation cover the most diverse 
fields. Among them is one on ** Forms of Liquids," one on the ''Veloc- 
ity of Light," one on the " Fundamental Concepts of Electrostatics." 
But perhaps the most valuable is that on the " Principle of the Conser- 
vation of Energy." The rise of this idea is traced in the history of 
mechanics, heat and electricity. The unity of these subjects is partic- 
ularly emphasized by a happy guess as to what might have been the 
result upon our electrical ideas had Riess devised his thermo-electrom* 
eter before Coulomb used his electrostatic balance. The chapter on 
the " Economical Nature of Physics" contains a discussion of the ques- 
tion as to what is meant by "a physical explanation," following in 



268 REVIEWS 

thought the now famous definition of Mechanics which Kirchhoff has 
given. One is apt to imagine now and then that the author is tread- 
ing dangerously near metaphysical ground ; but the chances are that 
the reader will find, ver}' shortly, that the only metaphysics involved 
is that which has crept into his own everyday thinking. A vein of 
humor is met here and there reminding the reader of Heaviside, never 
offending one's taste. These features, together with the lightness of 
touch with which Mr. McCormack has rendered them, make the volume 
one that may be fairly called rare. The spirit of the author is pre- 
served in such attractive, really delightful, English that one is assured 
nothing has been lost by translation. H. C. 



The Source and Mode of Solar Energy titroughaut the Universe, I. 
W. Heysinger, M.A., M.D. (J. B. Lippincott & Co.. 
Philadelphia, 1895.) 

This book must be relegated to the paradox shelf, although many 
parts of it are well written, many of its statements are true, and it is 
printed and bound in an excellent manner by a prominent publishing 
house, its author evidently has a considerable acquaintance with pop- 
ular scientific literature, and the facility with which well-known author- 
ities are quoted in apparent support of heretical views, together with 
the attractive appearance just referred to, make the book a dangerous 
one to the general reader. 

The author weighs the various theories which have been advanced 
to account for the maintenance of solar energy, and finds them all 
wanting ; he then proceeds to unfold a theory of his own. He is struck 
by the circumstance that free oxygen exists in the atmospheres of the 
Earth and other planets, but no free hydrogen, while in the atmosphere 
of the Sun there is free hydrogen but no oxygen. (He mistakes the 
absence of oxygen lines from the solar spectrum for proof that oxygen 
does not exist in the Sun.) This is for him the main fact to be consid- 
ered in any explanation of solar phenomena. The suggestion that the 
hydrogen molecules escape from the planetary atmospheres by virtue 
of their high velocity is dismissed in a few words, as unworthy of seri* 
ous consideration. Some agent is needed which will liberate oxygen 
in the atmospheres of the planets, and hydrogen in the atmosphere of 
the Sun, accompanied, in the latter case, by the production of light and 



REVIEWS 269 

heat. It is needless to say that this agent is electricity. Space is filled 
with attenuated aqueous vapor, the conductivity of which is doubtless 
increased by the presence of other gases. By the rotation of the 
planets powerful electric currents are generated, which, passing in cur- 
rents to the Sun, liberate hydrogen there by electrolytic action, the 
oxygen remaining on the planets. No sun can therefore exist without 
at least one planet ; each is the opposite and necessary complement of 
the other, and thus is established the dual relation, exemplified in the 
sexes, so dear to the heart of the paradoxer. 

The source of solar energy is after all not very clear, since the 
author objects strenuously to the supposition that the axial rotation of 
the planets is in any way affected. This is the more remarkable since 
the contradiction of the principle of the conservation of energy in Sie- 
men's hypothesis is clearly pointed out. In the long explanation of 
this subject in chapter IV the meaning is lost in words ; but the idea 
seems to be that the true source of solar energy is the ''potential energy 
of space," which is inexhaustible, and which is made active by the rota- 
tion of the planets, while rotation is an inherent property of matter, or 
rather, its persistence in the case of the planets, under the circum- 
stances considered by the theory, no more requires explanation than 
its original existence. It is needless to follow the detailed application 
of the theory to the various phenomena of the solar system, which it 
explains in a manner satisfactory to the author. J. E. K. 



Recent Publications. 



A Lisr of the titles of recent publications on astrophysical and 
allied subjects will be printed in each number of The Astrophysical 
Journal. In order that these bibliographies may be as complete as 
possible, authors are requested to send copies of their papers to both 
Editors. 

For convenience of reference, the titles are classified in thirteen 
sections. 

I. The Sun. 

EvERSHED, J. The Electric Origin of the Solar Chromosphere. Knowl. 

z8, 39, February, 1895. 
HowLETT, Rev. F. Remarks on Three Volumes of Sun Spot Drawings 

presented to the Society. M. N. 55, 73, 1894. 
QuiMBY, A. W. Sun Spot Observations. A. J. No. 334, 14, 175. 
Tacchini, p. R^sum^ des observations solaires, faites k Tobservatoire 

Royal du College Romain pendant les 2«, 3' et 4* trimcstrcs, 1894. 

C. R. Z20, 143-144. 
Tacchini, p. and Ricc6, A. Immagini spettroscopiche del bordo solare 

osservate a Roma e Catania nei mesi di novembre e dicembre del 

1893. (Plate) Mem. Spettr. Ital. 93, November, 1894. 

3. Stars and Stellar Photometry. 

Backhouse, T. W. Variable Orange Stars. Obs'y x8, 94, January, 1895. 
De Ball, L. Anzeige eines neuen verSnderlichen Stems BD. — 

6°. 5416 in Aquila. A. N. 137, 73-74. 
Bi&LOPOLSKY, A. l^tude sur le spectre de I'^toile variable S Cephei. 

Bui. I'Acad. Imp. de Sci. St. Petersburg, 267-306, November, 1894. 
Deslandres, H. Sur la vitesse radeale de { Hercule. C. R. zao, 1253. 
Fleming, M. Stars Having Peculiar Spectra. A. N. 137, 71-74. 
Maunder, E. W. Dark " Lanes " of the Milky Way. Knowl. x8, 

36-38, February, 1895. , 

Sawyer, Edwin F. On the Light- Variations of 7149 S Sagittae. A. J. 

No. 336, 14, 1 89-191. 
Tiss&rand, F. Sur T^toilc variable j8 de Persei (Algol). C. R. lao, 

125-130. 

Yendell, Paul S. On the Variable Star 103 T Andromedae. A. J. No. 

335.14, 183-184. 

270 



RECENT PUBLIC A TIONS 27 1 

5. Planets, Satellites and their Spectra. 

Brenner, L. Mars-Beobachtungen an der Manora-Sternwarte vor der 

Opposition 1894. A. N. 137, 49-62. 
Elger, T. G. Selenographical Notes. Obs'y xS, 85-86, January, 189$. 
Grossmann, £. Beobachtungen des Mondkraters Mdsting A und der 

Mondsteme. A. N. 137, 1 13-128. 
Johnson, Charles W. L. Pseudo-Satellites of Jupiter in the Seventeenth 

Century. Nat. 51, 285-286. 
Meunier, S. Recherches sur les conditions qui ont d^termin^ les car- 

act^res principaux de la surface lunaire. C. R. xao, 225-227. 
PRINZ, W. Le nouveau crat^re pr^s de Chladni et la limite de defini- 
tion des photographies lunaires actuelles. A. N. 137, 91-92. 
Schiaparelli, G. Alcune mutazione importanti osservate nella super- 

ficie di Marte. A. N. 137, 97-100. 
Schiaparelli, G. SuUe maree prodotte in un pianeta od in un satellite 

dairazione del suo corpo centrale. Atti. dell* I. R. Accademia degli 

Agiati, 1894, II pp. 

6. Comets, Meteors and their Spectra. 

Denning, W. F. Meteors of 1895, January 1. Obs*y 18, 93-94, Jan- 
uary, 1895. 

8. Terrestrial Physics. 

Bauer, L. A. Beitrage zur Kenntniss des Wesens der Sacular- Varia- 
tion des Erdmagnetismus. Mayer & Mikller, Berlin, 1895, 54 pp. 

Herschel, a. S. Aurora of November 23, 1894. Nat. 51, 246-247. 

Kitto, E. Report of Magnetical Observations at Falmouth Observatory 
for the Year 1893. Proc. R. S. 56, 557-563. 

Muller, p. J. Erdmagnetismus und Luftelektrizit&t. Htm. u. Erde. 
7» 153-167. 1895. 

MouREAUX, Th. Sur la valeur absolue des ^l^ments magn^tiques au i^ 
Janvier, 1895. C. R. lao, 42-43. 

9. Experimental and Theoretical Physics. 

Eder, J. M. und Valenta, E. Absorptionsspectren von farblosen und 

gefarbten Glasem mit Berttcksichtigung des Ultraviolett. Denkschr. 

d. Wiener Akad. 6x, 1 1 pp. 
FousSEREAU, G. Sur Tentrainement des ondes lumineuses par la mati^re 

en mouvement. C. R. xao, 85-88. 
GuiLLAUME, Ch. Ed. Application du prlncipe de Doppler k T^nergie 

des radiations. Jour, de Phys. 4 (3"« s^r.) 24-29, January, 1895. 



272 RECENT PUBLICA TJONS 

Olszeweski, K. The Liquefaction and Solidification of Argon. Chem 

News, 71, 59-60. 
Thomson, J. J. Electric Discharge through Gases. Nat. 51, 330-333. 

10. The Spectra of the Elements. 

Crookes, W. On the Spectra of Argon. (Abstract.) Chem. News, 71, 

58-59. 
Eder, J. M. und Valenta, £. Ueber die verschiedenen Spectren des 

Quecksilbers. Denkschr. d. Wiener Akad, 61, 30 pp. 
Eder, J. M. und Valenta, E. Ueber das Spectrum des Kaliums. 

Natriums und Cadmiums bei verschiedenen Temperaturen. 

Denkschr. d.- Wiener Akad. 61, 20 pp. 
EwAN, T. On the Absorption Spectra of Dilute Solutions. Proc. R. S. 

57, 117-161. 

11. Photography. 

BoNACiNi, C. La cromofotografia interferenziale e un nuovo caso di 
sensibilit^L pei colori. Mem. Spettr. Ital. 23, 146-154, 1894. 

AuGusTE et Louis Lumi^re. Sur les d^veloppateurs organiques de 
I'image latente photographique. Ann. Chim. et Phys. 4 (7 "^ s^r.) 
271-288, January, 1895. 

12. Instruments and Apparatus. 

Schuster, Arthur. Electrical Notes. Phil. Mag. February, 1895. 
Turner, H. H. Note on the Geometrical Optics of the Skew Casse- 
grain. Obs*y 18, 81-83, January, 1895. 

13. General Articles, Memoirs and Serial Publications. 

Larmor, J. The Significance of Wiener's Localization of the Photo- 
graphic Action of Stationary Light Waves. Phil. Mag. January, 
1895. 

MoNCK, W. H. S. The Structure of the Universe. Knowl. 18, 38, 
Februar)', 1895. 

Preston, S. Charles. Comparative Review of Some Dynamical 
Theories of Gravitation. Phil. Mag. February, 1895. 

Rayleigh, Lord and Ramsay, W. Argon : A new Constituent of the 
Atmosphere. (Abstract.) Chem. News, 71, 51-58. 

Rowland. Modem Theories as to Electricity. Engineering Mag. 
January, 1895. 

Tutton. a. E. a New Element in the Nitrogen Group. Nat. 51, 258. 



THE 

ASTROPHYSICAL JOURNAL 

AN INTERNATIONAL REVIEW OF SPECTROSCOPY 
AND ASTRONOMICAL PHYSICS 



VOLUME 



APRIL 189^ NUMBER 4 



RECENT RESEARCHES ON THE SPECTRA OF THE 

PLANETS. II.' 

By H. C. VoGEL. 
JUPITER. 

A PHOTOGRAPH of the spectrum of Jupiter was obtained with 
Spectrograph III on October 22, 1891. It extends from X487^^ 
to X 380 /Aft, but only fifteen Fraunhofer lines are certainly visible 
in this interval, as the slit was rather wide. No deviation from 
the solar spectrum can be detected. On October 24 of the same 
year an excellent photograph was obtained with Spectrograph II. 
The spectrum can be traced from F as far as H. It is very 
weak from F to X 446/4fi, and only a few lines can be seen in this 
part; the violet end from X405^^ to H is also quite weak. The 
remaining portion of the spectrum is, however, very rich in lines ; 
over one hundred lines in the spectrum of Jupiter can be identi- 
fied with lines in the solar spectrum. I give below the observa- 
tions of a small part of the spectrum which was specially investi- 
gated : 

> Continued from page 209. 

273 



274 




H, C\ VOGEL 


X 
4132.3 


Strong line 
Strong line 


X 
4204.0 


Line 


4134.8 


4207.0 


Line 


4I37.0 


Line 


4210.5 


Line 


4I40.2 


Line 


4216.0 


Strong line 


4142.3 


Line 


4217.6 


Line 


4i44.o 


Very strong line 


4219.5 


Line 


4148.0 


Line 


4222.5 


Line 


4150.0 


Narrow band 


4224.8 


Narrow band 


4152.2 


Line 


4227.0 


Very strong line 


4154.5 , 
4156.8 i 


System of lines 


4229.9 
4233.5 


Line 
Line 


4159.1 


Line 


4236.0 


Line 


4161.5 
4164.0 


Line 
Line 


4238.2 1 
4240.0 ' 


\ System of lines 


4165.8 


Delicate line 


4243.0 


Narrow band 


4167.7 


Strong line 


4246.3 


Band 


4173.0 


Narrow band 


4250.5 


Narrow band, diffuse toward 


4178.0 


Line 




violet 


4182.0 


Line 


4255.0 


Narrow band 


4185.0 


Line 


4260.5 


Strong line 


4187.5 


Very strong line / 
Strong line 1 


4264.3 


Weak band 


4192.0 


4268.0 


Narrow band 


4195.6 


Line 


4271.8 


Strong line 


4198.8 


Very .sharp line ) 






4200.1 


Distinct, sharp I 






4202.0 


Sharp line ) 







In the photographs that have just been mentioned the slit of 
the spectrograph was placed parallel to the direction of the diur- 
nal motion, and therefore nearly parallel to the belts ; but in four 
photographs taken on November i, 1894, with apparatus IV the 
slit was placed at right angles to the belts, in order to detect any 
differences of intensity between the spectrum of the strongly 
colored red belts and the spectra of other parts of the disk. Of 
the very successful photographs which were obtained, one, taken 
with 2" exposure, shows about seventy lines coinciding with 
solar lines ; the three others (two with 3" each and one with 4" 
exposure) show sixty-six lines. The two equatorial belts are 
very distinctly marked ; they appear on the negative as bright 
stripes extending throughout the length of the spectrum. The 



THE SPECTRA OF THE PLANETS 275 

brightness of these stripes sensibly increases toward the violet, 
but between F and G the contrast between the spectrum of the 
equatorial belts and that of the adjacent parts of the disk is 
slight. No other difference than that of the stronger general 
absorption can be detected in the spectrum of the red belts. 

Two of the photographs of Huggins, taken in 1878, represent 
the spectrum of Jupiter with that of the sky on each side. In 
ont of the photographs the sky spectrum extends only a little 
beyond K, while the planetary spectrum is strong up to X 333^1^1 
and can be traced as far as X326/Afi. The weak sky spectrum 
allows a comparison of only the principal lines in the two spectra ; 
comparison with the atlas showed a complete coincidence of 
fifty-seven planetary and Fraunhofer lines. 

The second plate is not so good ; only fifteen or twenty lines 
can be certainly recognized, which coincide perfectly with lines 
in the spectrum of the sky background. 

A remarkable observation on the spectrum of Jupiter by 
Henry Draper led to the publication of an article entitled, ** On 
a Photograph of Jupiter's Spectrum showing Evidence of Intrinsic 
Light from that Planet."* On September 27, 1879, Draper 
obtained a photograph of the spectrum of Jupiter with the slit 
placed at right angles to the direction of the belts, and on 
the negative was a narrow stripe running lengthwise through 
the spectrum nearly in the middle. The brightness of this 
stripe was unequal ; from h to beyond H it was very much 
brighter (on the negative) than the spectrum of the neighbor- 
ing parts of the disk, but below h it became less bright, and in 
the region between G and F it was even darker than the adjacent 
spectrum. According to the illustration which accompanies the 
article, the increase of brightness at // was very abrupt, and the 
deepest shade was at F. Draper attributed the stripe which ran 
through the spectrum to the equatorial belt, and concluded that 
luminous material must have been present in that region. It 
was not at a sufficiently high temperature, in his opinion, to emit 
the more refrangible rays, but it was hot enough to emit the less 

«il/..^^. 40, 433(1880). 



276 H. C. VOGEL 

refrangible rays which, on account of the absorption of the solar 
rays in the equatorial regions of Jupiter, appeared bright by con- 
trast in the lower spectrum. Draper's attention was not attracted 
by this peculiar spectrum until some time afterward, and he then 
found that the great red spot on the southern hemisphere, which 
at that time was the subject of much discussion, was on the 
middle of the disk when the photograph was taken. He there- 
fore attributed the peculiar spectrum to this spot. ''It may be 
that eruptions of heated g^ses and vapors of various composi- 
tion, color and intensity of incandescence are taking place on 
the great planet ; and a spot which would not be especially con- 
spicuous from its tint to the eye might readily modify the spec- 
trum in the manner spoken of above."' 

Now, an observation of Jupiter was made on September 26, 
1879, by Messrs. R. Copeland and J. G. Lohse at the Dun Echt 
Observatory, at a time when the red spot was also on the middle 
of the disk.* These observers, like Draper, placed the slit at 
right angles to the belts, in order to investigate their spectrum. 
They found that it was in exact agreement with my earlier 
observations on the belts ; ^ a dark stripe extended through the 
spectrum from the extreme red up to X 45 3^/*, the darkest part 
being in the vicinity of F. When the observer^, after having 
observed the spectrum of the satellites, again investigated dif- 
ferent parts of the disk of Jupiter, they found that the red spot 
on the southern hemisphere, which was then central, produced 
the same darkening of the spectrum as the equatorial belts. 
They add the following remark : "The absorption seemed to be 
more restricted to the region of spectrum near b and F than in 
previous observation." 

This observation cannot be brought into good agreement 
with what is shown on the plate taken by Draper. The spectrum 
of the spot between G and F could not possibly be brighter than 
neighboring regions (darker on the negative) if in the visual 
spectrum the effect of the spot was such as to cause a darkening 

«Jf.A^. 40, 435(1880). 
^Ibid,, p. 87*. 
^Ibid,, p. 88. 



THE SPECTRA OF THE PLANETS 277 

even in the "extreme red." It can only be assumed in expla- 
nation that on the following day some quite peculiar eruption 
from the interior of the pUnet had taken place, causing the 
spot to appear by direct observation not darker than the other 
parts of the disk, as it usually does, but brighter. But this 
assumption is contradicted by an observation made by Dr. Lohse 
at Potsdam, according to which the spot, although intensely red 
on September 27, 1879, presented in other respects the same 
appearance as on other days. 

Hence there remains only the assumption, the probability of 
which is very much strengthened by the appearance of the illus- 
tration, that the cause pf the peculiarity was a defect in the 
photographic film of the negative. 

I have made repeated observations of the red spot — which 
in the years 1880- 1883 was a remarkable object, awakening the 
most widespread interest — but I have never been able to find 
the least difference between its spectrum and that of the red 
belts. 

With regard to the satellites of Jupiter, the spectra of which 
I had investigated in my early observations, there is only one 
observation in recent times to be mentioned. It was made at 
Dun Echt by Copeland and Lohse,' who were, however, unable 
to see any lines in the continuous spectrum. Their remark that 
a dark longitudinal stripe could be seen at times when the spec- 
trum of the third satellite was greatly widened by means of a 
cylindrical lens, as if the satellite had an equatorial belt, I will 
mention for the sake of completeness only. No weight can be 
attached to the observation, since such a belt, if there were any, 
would not be visible if a cylindrical lens were employed. An 
appearance similar to that observed is also produced when the 
spectrum of a star is greatly widened. 

On the strength of my early observations I ventured the 
statement that the lines in the red, which are so characteristic 
of the spectrum of Jupiter, were probably also present in the 
spectra of the satellites ; this would indicate that the satellites 

>Af.iV.4o,88. 



278 H. C. VOGEL 

are surrounded by atmospheres similar to that of the primary. 
So far, however, no observations appear to have been made 
elsewhere on this subject, although with large instruments they 
could be made with some prospect of success. 

Photographs of the spectra of the satellites have been made 
at Potsdam. On November 25 and 26 Mr. Wilsing obtained 
photographs of the spectrum of satellite III. The exposures 
were 20" and 30". On the first plate the spectrum extends from 
X 487^^11 to X370/i/i; it is very strong, in the neighborhood of G 
even somewhat over-exposed, and shows forty Fraunhofer lines. 
The second plate shows forty-four lines. A number of lines 
also appear on photographs of the spectra of satellites I, II and 
IV, which were made on January 14, 1895. 

SATURN. 

Only one photograph of the spectrum of Saturn has been 
made at Potsdam. It was obtained on March 17, 1892, with 
1 5"* exposure. The spectrum, which extends from F to Ht, is 
strong, and about thirty Fraunhofer lines are visible in it. The 
slit was rather wide, in consequence of which the fine lines are 
blended together into groups or bands. Narrow spaces in the 
spectrum near Hy and G, and at several other places where there 
is an absence of lines, appear strikingly bright, and at first sight 
produce the effect of bright lines ; but a spectrum of the Moon 
taken on March 5, 1892, with the same adjustment of the slit 
presents exactly the same appearance. No deviation from the 
solar spectrum can be recognized. 

Two photographs were made by Huggins in 1887. One of 
them, taken on March 23, with an exposure of l\ is remarkable 
for the great extension of the spectrum into the ultra-violet ; it 
can be traced as far as X3 15^/11. The slit was wide, and therefore 
only twelve lines can be recognized. A photograph of less excel- 
lence, taken on March 19, with about 20° exposure, shows only 
the lines //y, G, H, K, and two groups of lines in the ultra-violet. 
I have also six photographs taken by Mr. and Mrs. Huggins in 
1889, which have an especial interest on account of the fact that 



THE SPECTRA OF THE PLANETS 279 

the slit was so placed as to give the spectrum of the ansae of the 
ring on each side of the spectrum of the ball. Not the least dif- 
ference can be perceived between the spectra, which, on three of 
the plates especially, are distinctly separated. 

Three of the six photographs contain little detail, and only 
the principal lines can be seen in the spectrum of the planet or in 
that of the sky background which accompanies it. A fourth plate 
without comparison spectrum shows twenty-two lines in the 
ultra-violet between H and X 344^1^1. The spectrum can be 
traced as far as A 3 30^1^1; altogether, more than thirty Fraun- 
hofer lines can be recognized. 'At Hy and H there are dark 
bands on the negative, which look as if they were caused by 
bright lines, as in the case of the Potsdam photograph. On 
the fifth plate they are wanting ; the slit was decidedly narrower. 
The spectrum extends only a short distance into the ultra-violet. 
Altogether twenty lines are visible in the planetary spectrum, the 
agreement of which with the solar spectrum is complete. On 
the sixth plate the lines which can be identified are Hy^ ^4325, 
G, Hh^ H, K, and more than twenty lines more refrangible than 
K. Three lines between F and G can be seen, but not F. All 
lines agree with lines of the bright sky background. 

In a note dated February 17, 1889, Mr. Lockyer* points out 
that in view of the hypothesis, which is becoming more and more 
firmly established, that the rings of Saturn are composed of small 
bodies of meteoric nature, it would be a matter of interest to 
investigate their spectrum. For, in case the collisions between 
the small bodies were accompanied by the evolution of gas and 
by the emission of light, it might be possible that some modifi- 
cation of the spectrum of the rings would be produced. He was 
led to this conjecture by the appearance of a photograph of Sat- 
urn taken by the Henry brothers, on which the brightness of the 
rings as compared with that of the disk of the planet was much 
greater than in visual observations. 

As I have already mentioned, the photographs of Huggins 
show definitely that there is no difference between the spectrum 

^A.N. 2881. 



280 H, C. VOGEL 

of the planet and that of the rings in the more refrangible parts 
of the spectrum. Lockyer's note, however, induced Mr. Keeler 
to investigate the spectrum of Saturn and its ring, at the Lick 
Observatory. His observations are published in A.N. 2927, and 
for the details of the investigation I may refer the reader to this 
excellent paper, in which the ideas of Lockyer are very com- 
pletely refuted, and in which it is above all pointed out that the 
first effect of any possible self-luminosity would be to make the 
rings visible in the shadow of the planet. 

In his investigation of the spectrum of the ring, Keeler was 
able to confirm an early observation of mine, which presented 
great difficulties on account of the small size of the image in the 
focus of the Bothkamp refractor. He found that in the spectrum 
of the ring there was no trace of the absorption band at X6I8flf^ 
which is so characteristic of the spectrum of Saturn. 

The greater brightness of the rings as compared with the 
planet, particularly for the chemically active rays, is quite natu- 
rally explained by the absence of an atmosphere around the ring 
and the extraordinary density of the atmosphere of the planet 
itself. 

URANUS. 

After several unsuccessful attempts Mr. Frost obtained a 
good photograph of the spectrum of Uranus with apparatus HI 
on April 23, 1892. The exposure was i** 20". The greatest 
intensity of the spectrum is between F and G, as the violet and 
ultra-violet rays were much weakened by absorption on account 
of the low altitude of the planet. The spectrum can, how- 
ever, be traced as far as H ; K cannot be recognized with 
certainty. Measurement established the presence of the fol- 
lowing lines, the corresponding solar lines being also given for 
comparison : 

SPECTRUM OP URANUS. SOLAR SPECTRUM. 

X X 

397: 3969 H 

4101 Distinct line 4102 Strong line surrounded by 

delicate lines ; h 



THE SPECTRA OF THE PLANETS 



281 



X 


SPECTRUM OF URANUS. 


X 


SOLAR SPECTRUM. 


4135 


Very weak 


4133 


Strong lines 


4145 


Very weak 


4144 


Strong double line 


4157 


Very weak and faint 


4155 


Group of lines 


4200 


Easily visible 


4200 


Group of rather strong lines 


4240 


Easily visible 


4227 


Strong line 


4260 


Narrow 


4260 


Double line in a group of 
delicate lines 


4275 


Broad and strong 


427s 


Line ) With low ditpenion. 






4272 


Strong double > h«ve the effect of a 


4288 


Narrow band 




line S •««>"- 


4291 


Narrow band 


4291 


Group of lines 


4300 


Narrow band 


4300 


Group of lines belonging to the 
G group 


4310 


Broad diffuse band 


4308 


Broad G group 


4320 


Conspicuous dark space on 
negative 


the 4317 


Nearly vacant region 


4325 


Broad line 


4325 


Strong line in a group 


4341 


Narrow band 


4341 


Strong line H*i and group 


4383 


Narrow band 


4384 


Very strong line 


444: 


Rather broad band 


4443 


Several lines 


446: 


Rather broad band 


4458 


Extended group of lines 



A successful photograph of the spectrum of Uranus made by 
Mr. and Mrs. Huggins on June 3, 1889, with 2" exposure/ 
extends farther into the ultra-violet than the one made at 
Potsdam. On this plate the maximum intensity is at G. There 
is a very beautiful sky spectrum, full of detail, on each side 
of the planetary spectrum, obtained by an exposure on the 
following morning with a considerably narrower slit. Although 
no direct comparison can be made with the planetary spec- 
trum on this account, very desirable reference points are 
obtained. 

The measurements of these plates which I have made give 
the following wave-lengths of the lines, or of the bands caused 
by blending of groups of lines in consequence of the rather wide 
slit. Here also I have added the wave-lengths of the corre- 
sponding solar lines : 

«5fe. (2 hours ?).— Ed. 



282 


//. C. 


VOGEL 


\ 


SPECTRUM OP URANUS. 


\ 


SOLAR SPECTRUM. 


3565 


Band 

Broad band { Scarcely visible 


3567 


Group of lines 


3580 


3584 


Group of broad lines 


363: 


Band 


3625 


Middle of a broad band consist- 
ing of several systems of lines 


3725^ 




3722 


Group of strong lines 


3737 
3750J 


' Lines, just visible 


3736 


Group of strong lines 




3748 


Group of strong lines 


3795 


Band 


3795 


Group of strong lines 


3833 


Broad band 


3833 


Group of strong lines 


3880 


Weak band 


3879 


Strong line in a group of rather 
strong lines 


3935 


Broad bright band (on nega- 
tive) 
Broad bright band (on nega- 


3934 


K 


3972 


3969 


H 




tive) 






4101 


Distinct line 


4102 


Strong line ; A 


4305 


Broad band, diffuse toward 


4305 


Middle of the group G 




violet 






4322 




4325 


Strong line in a j;roup 


4340 




4341 


Strong line //y and group 


486 




4862 


F 



The two photographs taken in Potsdam and in London sup- 
plement each other admirably, and together furnish a proof that 
the more refrangible portion of the spectrum of Uranus contains 
neither absorption bands nor bright lines ; hence the assertion of 
Lockyer' that the spectrum of Uranus is to be regarded as 
an emission spectrum is entirely without foundation. 

The careful investigation of the visual spectrum which was 
made by Mr. Keeler* with the 36-inch refractor of the Lick 
Observatory is in perfect accordance with this result. He 
expressly mentions that when he first looked at the spectrum 
the brightness of certain places in the yellow and green with low 
dispersion was such as to produce the effect of self-luminosity ; 
but further investigation with different spectroscopes convinced 
him that this impression was illusory, and due to the contrast 
of the bright places in the continuous spectrum with the neigh- 
boring dark absorption bands. 

' A, N, 2904 « A, N. 2927. 



THE SPECTRA OF THE PLANETS 



283 



Mr. Keeler was able to see the continuous spectra of the two 
outer satellites of Uranus. 

I extract from this paper the description of the spectrum of 
Uranus, and give for comparison an abstract of my earlier obser- 
vations, omitting from the latter some rather uncertain details. 
To the very excellent representation of the visual spectrum which 
accompanies Keeler's article I desire to call special attention. 
It corresponds perfectly with the appearance in the spectroscope, 
and is in so far to be preferred to the drawing given in my oft- 
mentioned memoir on planetary spectra that it takes into account 
the compression of the prismatic spectrum toward the red, and 
therefore more nearly represents the appearance with a prism 
spectroscope. Considering the extreme difficulty of the object 
the agreement of the drawings, as well as that of the following 
measurements, is very good : 



X 

6182 



6085 
5962 



5868 

5768 



KEELER. 

Middle of an absorption band 
which seems to be the strongest in 
the spectrum. With the narrow- 
est slit that could be used the 
breadth of this band was but little 
altered, showing that it is really 
a broaa band like the correspond- 
ing bands in the spectra of Jupiter 
and Saturn. 

(Brightest place in the red.) 
Dark absorption band which nar- 
rows with the slit and is therefore 
a comparatively sharp line. 
(Brightest part of the spectrum, in 
the yellow.) 

Darkest part of a broad absorption 
band with ill-defined borders. The 
middle of the band is farther up in 
the spectrum, at about X575AIM, 
and at this place there is a slight 
increase in brightness. The band 
has the appearance of being made 
up of several smaller ones. It is 
better defined on the lower than on 
the upper side. 



X 

6180 



VOGEL. 

Darkest place of an absorption 
band the breadth of which is from 
5 to 6/1^1. (Somewhat less con- 
spicuous than the band at X 543 ^.) 



5961 Rather weak, narrow band. 



5738 Darkest place of an absorption 
band which is diffuse on both sides, 
although somewhat more definitely 
bounded toward the red. The 
breadth is i o aim* (Not quite so dark 
as the band at X6i8fMA.) 

Remark, — According to Keeler's 
measurement the band is identical 
with the telluric group d. 



284 



H. C. VOGEL 



X 

564: 
552 



5425 



518: 



509: 



4850 



KEELER. 

(Bright place in the green). 
(Another bright place in the green. 
Between these two places there is 
a faint shade). 

Middle of a great absorption band 
almost as dark as that at X6l8fMi, 
but somewhat broader, and with 
edges not so well defined. 
Very faint band; position esti- 
mated. Perhaps the b group of 
the solar spectrum. 
Another very faint band ; position 
estimated. 

Decided band, but dim with high 
dispersion. Probably at the place 
of F (X4862) but too strong for a 
solar line, and without doubt due 
to absorption in the atmosphere of 
Uranus. 



X 

558: 



5425 



VOGEL. 



Weak line. 



Middle of a broad absorption band 
the width of which is about 5m#k; 
somewhat di£Fuse towanl the violet. 
The most conspicuous absorption 
band in the spectrum. 



508: Weak band. 

4861 Band, or broad diffuse band. 



ON THE PERIODIC CHANGES OF THE VARIABLE 
STAR Z HERCULIS. 

By N. C. DuNER. 

Toward the end of July, 1894, Mr. Chandler discovered that 
the star BD. + IS^'SSH, whose place for 1900.0 is 

«=i7'53"36*.o6, 8= + i5°8' 47'.2, 

is a variable of the Algol type, and he found its period to be 
3** 23"* 50". Circumstances prevented him, however, from observ- 
ing the star except when its light was increasing, and he there- 
fore communicated his discovery to a number of European 
astronomers, among them the writer, in order that if possible 
more complete observations might be obtained. I received Mr. 
Chandler's ephemeris on September 12, but on account of 
unfavorable weather I was unable to observe a minimum until 
September 18. In the mean time (on the 15th) a telegram 
arrived from Professor Hartwig, announcing that he had dis- 
covered the variability of the same star, and that its period was 
I** 23** 55" 40*. This simultaneous, independent discovery by two 
different persons would certainly have been ver}' surprising, if 
both of them had not mentioned that their attention had been 
directed to this star by a note in the admirable Photometric 
Durchmusterung of Messrs. Miiller and Kempf, in which it is 
stated (p. 482) that the star is of a suspicious character and 
requires further watching. 

In consequence of Professor Hartwig's telegram I resumed 
observation of the star on September 20, when the sky was again 
clear and, in fact, determined a minimum epoch. But while 
the star sank on the i8th more than a whole magnitude below 
its usual brightness, on the 20th the reduction of light at mini- 
mum was not half a magnitude ; and instead of occurring a few 
minutes earlier than on the i8th, as the information received 
from both discoverers led me to expect, the minimum was 

285 



286 iV. C DUNER 

early by a whole hour. On the strength of this observation 
I sent the following telegram to Professor Kriiger on Febru- 
ary 21 : 

"The new variable is probably of the Y Cygni type, with 
unequally bright components. Faint and very bright minima 
alternate; periods 47 and 49 hours." 

Further observations have shown, as I will explain fully 
below, that these words still represent the facts, as far as they are 
known, with very considerable exactness. 

On September 20 and 21 Professor Hartwig wrote to Pro- 
fessor Kriiger with reference to his discovery of the variability 
of the star and his researches on its periodic changes. Hart- 
wig, like myself, found an analogy with Y Cygni, but it was 
evident that he had not obtained complete observations of 
the uneven minima, and hence regarded the analogy as per- 
fect ; i, e,, he believed the components to be equally bright, 
and hence was led to assume that the uneven minima occurred 
52 hours before the even ones. My observations of September 
20 and 24 show, however, that my views as stated above are 
correct. On all three days at the beginning of observation 
(1. e., fully three hours before daylight, when the even minima 
occurred) I have seen Z Herculis brighter than BD. + M'^SSZS 
and found that its light was diminishing toward a minimum two 
hours later. If the secondary minimum occurs four hours 
earlier than the principal minimum, Z Herculis should have been 
only a little brighter than BD. + 15° 3301, or perhaps about as 
bright as BD. + 15^*3309, at the time when my observations 
began. 

Some time after this Herr Lindemann, of Pulkowa, announced 
his opinion that the principal minimum is double, so that two 
minima, separated by a minimum of less pronounced character, 
occur in the course of an hour. I have carefully studied the 
observations of Herr Lindemann, but I cannot share his opinion, 
for the quite numerous observations which have been made by 
several European astronomers according to Argelander's method 
do not show any such depression in the light curve, and Linde- 



PERIODIC CHANGES OF Z HERCULIS 



287 



mann's observations themselves by no means support this hypoth- 
esis in an unqualified manner. Thus, if simple curves of the 
ordinary form are drawn, it will be found that they represent all 
of the two days' observations quite well, with the exception of a 
single measurement (the next to the last on the second day), 
which shows a discrepancy of 0.4 magnitude. No other obser 
vation di£Eers from the curve by more than 0.2 magnitude, and 
even such discrepancies as this occur very rarely and are not at 
all systematically distributed. 

The minima which I have myself been able to observe are as 
follows : 



Epoch 


Minimum, Gieenwich M. T. 


Brightness 


Remarks 





1^1^ Sept. i8<* S** 44" 


7.0 


Good 


I 


" 20 7 18 


14.0 


Very good 


3 


" 24 7 5 


15.0 


Good 


6 


" 30 7 59 


6.0 


Very good 


8 


Oct 4 8 32 


4.6 


Good 


12 


" 12 7 54 


.... 


Quite uncertain 


18 


" 24 7 9 


4.0 


Rather uncertain 


36 


Nov. 29 5 22 


4.0 


Fairly certain 


37 


Dec. I 4 33 


13.8 


Quite good 



The extremely bad weather in October and November inter- 
fered very greatly with the observations. 

In order to derive elements from all the observations so far 
published, I first of all determined the following approximate 
elements from my own observations of the even epochs : 

E 

Min. = 1894*0 -f- 261^37 + 3.992 — . 

2 

I then expressed all observations of the even epochs in 
days and fractions of a day, reduced them to the Sun and 
compared them with the above elements. In doing this I was, 
however, unable to avail myself of Chandler's observations, as 
they have not yet been published. The following results were 
obtained : 



288 



N. C. DVNtR 



E. 


Gi«enwich M. T. 


Obt^— Coinp. 


ObMrver 


— 2 




2 


257**. 372 
261 .342 
261 .360 
261 .360 
261 .364 
265 .355 


1 M 1 1 t 
q 


H. 

H. 

Pann. 

P^ 

Pann. 


4 
6 
6 
6 
6 
8 


269 .354 
273 .348 
273 -332 
273 .333 
273 .323 
277 -355 


.000 
+ .002 

— .014 

— .013 

— .023 
+ .017 


H. 
H. 
D. 
PI. 
Li. 
D. 


12 

i6 
i8 


285 .327 
293 .279 
297 .296 


+ .005 

— .027 

— .002 


D. 
Li. 
D. 


36 


333 .220 


— .006 


D. 



In the column ** Observer," D. stands for Dun^r, H. for Hart- 
wig, Li. for Lindemann, Pann. for Pannekoek, PI. for Plassmann, 
and Prag for the observers Gruss and Laska at that place. The 
above differences between observation and computation have been 
combined to form normal differences in the manner indicated by 
the horizontal lines ; but in so doing Hartwig's second observa- 
tion was rejected, as Hartwig himself notes that the atmospheric 
conditions on that day were extremely bad. No considerable 
effect on the result would be produced, however, if this observa- 
tion were included. I have given my observations of the 12th 
and 1 8th minima only one-fourth the weight of Lindemann's, 
since both were observed on only one side of the light curve, 
and the first, in particular, was very uncertain. In this way the 
following equations of condition were obtained : 

X = — 0.008 

* + SJ' = ~ o.ooi 

jp + %y -=, — 0.017 

X •\- \%y = — 0.006 
From these equations I obtained the following corrected 
elements : 



Even Epochs, Min. = 1894.0 + 26i**.36i -}- 3^99201 



R 



PERIODIC CHANGES OF Z HERCUUS 
The residual errors are exhibited below : 



289 



£. 


OU.-<:omi». 


Ob«mr 


E. 


Ob«.~Comp. 


OoKnrcf 


— 2 


+ 0«».0O3 


H. 


6 


— tf*.oos 


D. 





— .019 


H. 


6 


— .004 


PI. 





— .001 


Pann. 


6 


— .014 


Li. 





— .001 


Prag 


8 


--0 .026 
--0 .014 


D. 





- 


-0 .003 


D. 


12 


D. 


2 


- 


- .002 


Pann. 


16 


- .018 


Li. 


4 


- 


-0 .009 


H. 


18 


+ .007 


D. 


6 


^ 


[-0 .oil 


H. 


36 


-ho .003 


D. 



Only three uneven minima have so far been observed, all of 
them by me. I have not used them for determining the period, 
partly on this account, and partly because the small amount of 
the whole change during these minima renders the determination 
of the epochs quite uncertain. The following elements are there- 
fore determined with reference to the period which was found 
for the even epochs : 

Uneven Epochs, Min.= 1894.0+ 263''.3i2+3''.992oi . 

The observations are thus represented : 

E. Min. Obs. — Comp. 
I 263''.304 — o**.oo8 
3 267 .295 — o .009 
37 335 -1^6 +0 .018 
From the two systems of elements I have computed the fol- 
lowing ephemeris : 

EPHEMERIS OF Z HERCULIS. 



E. 


Even Epochs. 






E. 




Unevoi Epodu. 






100 


1895, April 


5* 


23I1 


4« 


lOI 


X895. 


April 


7" 


2l'» 


45' 


120 


May 


15 


21 


10 


121 




May 


17 


19 


51 


140 


June 


24 


19 


IS 


141 




June 


26 


17 


56 


160 


August 


3 


17 


19 


161 




August 


5 


16 


I 


180 


September 


12 


15 


24 


181 




September 


14 


14 


6 


200 


October 


22 


13 


29 


201 




October 


24 


12 


10 



In the above ephemeris, as elsewhere in this article, Green- 
wich mean time is understood. 

It will be seen at once that the star cannot be observed in 
Europe in 1895. ^^ ^^ other hand, observations can be made 
under favorable circumstances in America — at the California 



290 N, C, DUNtR 

observatories as early as the beginning of spring. The star is 
one of the very greatest interest, and it is greatly to be wished 
that it may be diligently observed, all the more because in 1896 
observations can probably be made only in Asia and Australia. 

Here I might close this article ; but the temptation to enter 
into some speculations on the probable constitution of the sys- 
tem of bodies which compose the star is too strong to be resisted. 
Z Herculis occupies a unique position among the stars of the 
Algol type, inasmuch as its minima are of different degrees of 
brightness, regularly alternating between faint and very bright. 
As we now know that the interval between a bright and a faint 
minimum is greater than that between the latter and the next 
bright minimum, there can be no doubt that Z Herculis belongs 
to the Y Cygni type, and that the system does not, like Algol, 
consist of a bright and a dark body, but of two bodies, both of 
which are bright. But contrary to the case of Y Cygni, one com- 
ponent must be brighter than the other. The observations which 
are available suffice not only to determine the relative magni- 
tudes of the two components, but also their relative brightness 
per unit of surface. 

Herr Lindemann determined the magnitude of the star at its 
usual brightness and also at the time of a principal minimum, and 
found it to be At maximum = 6^.89, 

At principal minimum=8 .05, 
and with these data I find that the magnitude is 
At secondary minimum=7".35. 

The relative degrees of brightness at maximum, uneven mini- 
mum and even minimum, are as 

If we let ' = ^ = ^- 

y4 = the surface of the brighter star, 
^y^=the surface of the fainter star, 
i = the brightness of the unit surface of the first star, 
j^=the brightness of the unit surface of the second star, 
the total brightness of Z Herculis at maximum, uneven and even 
minima respectively, will be represented by the following equations: 



PERIODIC CHANGES OF Z HERCULIS 29 1 

A-^Axy = I, 

A =^, 

A'-Ax+Axy^yi. 

Subtracting the third from the first equation we obtain 

Ax=}i, 
and combining this with the second equation, 

X=: I, 

Hence Z Herculis consists of two components of equal size, 
one of which is twice as bright as the other. It is here assumed 
that the mutual eclipses are central, or nearly so ; still I have 
not been able to hit upon any other assumption that represents 
the observations satisfactorily. 

Still further conclusions can be drawn from my observations. 
Thus, observations extending through a whole period show that 
near the time of a principal minimum the variation of light 
extends over 6.6 hours, while near the time of a secondary mini- 
mum the entire variation requires only 4.0 hours. The orbits of 
the stars must therefore be considerably eccentric, and the bright 
minimum falls nearer to the perihelion, the faint minimum nearer 
to the aphelion. 

If we assume that the bright minimum falls exactly at peri- 
helion, and the faint minimum exactly at aphelion, we can deter- 
mine the eccentricity of the orbit ; for we have for perihelion and 

aphelion respectively, a cos^ 

r.av = dJIf, 

I —e 

or, r^dv^ __ 1+^ 

r^v^ I — e 

Now ' ' is the ratio of the arcs described by the two 

bodies in a unit of time, or, under the assumption which has been 
made, the ratio of the durations of eclipse at the even and uneven 
minima. Hence, 1 + ^ _ 6.6 

I — e^ 4.0 



292 N. C. DUNtR 

from which ^=0.245, this value of e being the lower limit of 
eccentricity. 

According to our elements the minima of the epochs o, i, 2 
occur at the following times : 

E. Min. Diff. 

I 263 .312 

a »6s .353 +" •°^' 

It is therefore clear that the minima do not fall exactly at the 
perihelion and aphelion points, as in that case the above differ- 
ences would be equal, but it is also clear that they must fall very 
close to these points. With the aid of the above differences the 
angle between the line of apsides and the line of sight can be 
found tentatively. After several trials I have assumed that 

^±:0.247S- 

As the line of sight must evidently pass through both stars at 
the epochs of both minima, the true anomalies must then be 
respectively ^^ and v.= i8o°+t^. 

Hence, giving v^ some hypothetical value, the mean anomalies 
M^ and M^ can be determined by the formula 

tani^ = tan^e^-Ji^, 

M-^E—e sin E, 
Now, since the interval between the epochs o and i is 1.95 1 days, 
and that between i and 2 is 2.041 days, it is evident that aphelion 
occurs during the latter interval. Hence M^ must be numerically 
greater than M^^ and v^ numerically greater than v^. We have 
then 2.0A1 

M.-- M,= -^ .360^ =: 184^06. 

I.95I + 2.041 -^ 

The assumed value v^ is correct if the values of 3/. and M^ found 
by it differ by 184^.06. After several trials I found : 

«'«=3''-98, »,= i83*'.98, 
^,=3 .09, ^.= 185 .02, 
M,-2 .33, i/'.= i86 .39, 

which exactly fulfil the required conditions. 



PERIODIC CHANGES OF Z HERCULIS 293 

Finally, we must ascertain whether the assumed value of e is 
sufficiently accurate. For this purpose we first compute, with 
the aid of the values of M^ and M^ just found and the durations 
of the two minima, the mean anomalies M[ , M[ respectively at 
the beginning and at the end of the even minima, and the corre- 
sponding values M[, M\ for the uneven minima. From these 
we deduce the corresponding true anomalies «//, t/', v[, v\, and 
finally the linear values, corresponding to these anomalies, of 
r[ sin {p-v[\ rl sin K -^), r,' sin (f .-t','), r\ sin K— f ,). 
Then, if the assumed e is correct, these last four values must be 
equal, since the stars are practically at an infinite distance. 
The durations of the minima found above were 
For even minima, 6** 36"= ©''.a 7 500, 
For uneven minima, 4 o =0 .16667. 

Hence .., ^^,_ 0.16667 ,^ o_,^o^, 

ilf. — ilf, = .360 =15.03, 

3.99201 



M\- 


j^,_ 0.27500 


.360": 


=24°. 


80; 


• 3.99201 


M[ = 


-5°.i8 


Ml 


= 9' 


•8s 


K = 


173-99 


K 


= 198 


•79 



from which we obtain : 

v[ = — 8*^.85, vl = + i6°.76, log r/ = 9.87753, log r,' = 9.88018, 

v[ = 176.26, vl = 191.76, log r; =0.09574, log r; =0.09306, 

r/ sin {v, — v^) = 0.1675, r,' sin (t^/ — v,) = 0.1679, 

r,' sin (f, — f,') = 0.1675, r; sin «—»,) = 0.1677. 

This agreement is far more complete than is necessary, since 
even the third decimal may be uncertain by several units. The 
eccentricity which we have found is somewhat uncertain in con- 
sequence of this fact, but an eccentricity less than 0.2 or greater 
than 0.3 does not seem to me to harmonize with my observations. 

The linear values derived above are evidently equal to the diam- 
eter of the stars expressed in terms of the major axis of the orbit. 

Collecting what has been ascertained in the preceding inves- 
tigation with respect to the constitution of Z Herculis, we have 
the following result : 



294 N. C. DUNiR 

Z Herculis consists of two stars of equal size, one of which is twice 
as bright as the other. These stars revolve around tlteir center of 
gravity in an elliptical orbit wliose semi-axis major is six times tlic 
diameter of the stars, ' The plane of the orbit passes through the Sun, tlie 
eccentricity is 0,2475, and t/ie line of apsides is inclined at an afigle 
of 4"" to the line ofsiglit. The stars revolve in this orbit in j days, 2j 
hours, 48 minutes, jo seconds. 

Hence Z Herculis stands in an isolated position among stars 
of the Algol type, or rather it forms a hitherto missing link 
between the stars of the pure Algol type and Y Cygni. As I 
have already remarked, it is highly desirable that the star should 
be diligently observed by American astronomers during the 
present year. The last part of the above investigation shows, 
moreover, that it is not enough to make a large number of obser- 
vations for determining the minimum epochs ; the duration of the 
light-changes is of equal importance and should be determined 
as sharply as possible. Hence, when circumstances permit, the 
observations should begin while the star is still at its full bright- 
ness, and continue until full brightness is again restored. Accurate 
photometric measurements shouhd also be made of the magni- 
tude at maximum, as well as at both minimum epochs. The 
above investigation shows how much respecting the nature of the 
star such observations would reveal ; at the same time I would 
be the first to acknowledge that the numerical data I have 
employed are greatly in need of revision, and that it will be neces- 
sary to wait for observations made under favorable conditions' 
before we can draw trustworthy conclusions as to the numerical 
relations of the system. These observations should moreover 
be made very soon. At present the angle between the line of 
apsides and the line of sight is small. But judging by the 
case of Y Cygni, we may expect that this angle is subject to 
rapid changes. No opportunity should therefore be neglected 
to determine this element as soon and as well as possible. 

' It is here assumed that one star remains fixed in the focus of the elHpsc. 
"The conditions were very unfavorable here in 1894. 



PRELIMINARY TABLE OF SOLAR SPECTRUM 
WAVE-LENGTHS. IV. 

By Henry A. Rowland. 







UlCBsitT 






iDlosity 


Wa^.ldvtli 


SnlMtaiioe 


and 
Qutfadcr 


Wmweltngth 


SubMaaoe 

1 


aad 
Chancier 


4266.233 




000 


4272.114 


1 
1 


IN 


4266.374 


Ti 


00 


4272.299 


1 


oooN 


4266.586 




OON 


4272.458 




oN 


4266.778 




' 


4272.590 




00 


4266.894 


Cr 





4272.701 


Ti- 


I 


4267.122 


Fe 


3 


4272.869 




000 


4267.291 




oooN 


4273049 




I d? 


4267^39 




000 


4273.274 




ooN 


4267-543 




2N 


4273.482 


S* 


3N 


4267.741 




00 


4273.643 


Zr 


2N 


4267.900 1 




I 


4273.835 




00 


(4267.950) ^s 






4273.946 







4267.985 J 


Fc 


3 


4274.045 


Fe? 


I 


4268.138 


Zr 





4274.096 




I 


4268.266 




I N 


4274.348 




2N 


4268.449 




oooN 


4274.542 




ooN 


4268.606 




oooN 


4274.746 


Ti 


2 


4268.783 







4274.958 s 


Cr 


7d? 


4268.915 


Fe 


2 


4275.115 







4269.088 




ooN 


4275.262 




00 


4269.202 




00 N 


4275.413 




oN 


4269.339 




000 


4275-541 




N 


4269.446 







4275.667 




0000 


4269.625 


La 





4275.713 







4269.746 




00 


4275.814 







4269.898 




2 


. 4275866 







4270.016 




2N 


4276.050 




000 N 


4270.120 




00 


4276.150 




ooN 


4270.329 


Ti- 


iN 


4276.259 







4270.485 




00 


4276.428 







4270.649 




ooN 


4276.587 


Ti 





4270.880 




ooN 


4276.685 




000 


4271. 112 




ooooN 


4276.836 


Zr 


2 


4271.212 




oN 


4276.979 




000 


4271.325 


Fe 


6 


4277-147 


V- 


I N 


4271.538 







4277.384 




I N 


4271.618 




00 


1 4277.544 


Zr 





4271-715 




00 


' 4277.692 




2d? 


4271.790 




oNd? 


4277.835 




00 


4271.934 s 


Fe 


15 


] 4277.968 




0000 



295 



296 



HENRY A, ROWLAND 







Intesfti^ 






lolensitj 


Wave-length 


Substance 


end 
Character 


Wave-length 


SubManoe 


and 
Character 


4278.060 




0000 


4285.605 


Fe 


3 


4278.153 




0000 


4285.692 




I 


4278.308 







4285.834 




000 


4278.390 


Fe-Ti 


3 


4285.966 




I 


4278.595 




00 N 


4286.092 




00 


4278.704 




0000 


4286.168 


Ti- 


2 


4278.843 




00 


4286.244 







4279.009 


Ti- 


xN 


4286.350 




I 


4279.225 




IN 


4286.478 




000 


4279.380 




000 N 


4286.627 




3N 


4279.475 




000 N 


4286.741 




00 


4279.643 




2 


4286.893 




000 


4279.874 




2Nd? 


4287.034 




I 


4280.027 




I 


4287.159 


La 


2 


4280.194 




I 


4287.205 







4280.374 




X 


4287.394 




ooNd? 


4280.494 




000 


4287.566 


Ti 


I 


4280.556 


er 


I 


4287.736 




00 


4280.647 




000 


4287.873 




000 


4280.698 







4288.038 


Ti 


2 


4280.789 







4288.149 


Ni 


X 


4280.938 




I 


4288.310 


Ti 


I 


4281.1x3 




I 


4288.423 




0000 


4281.257 


Mn 


2 


4288.562 




ooNd? 


4281.410 




000 


4288.721 




0000 


4281.530 


Ti 





4288.888 




ooN 


4281.648 




000 


4289.115 




I 


4281.752 




00 


4289.237 


Ti 


2 


4281.903 




000 Nd? 


4289.365 




0000 


4282.127 




2N 


4289.525 s 


Ca 


4 


4282.370 




00 


4289.695 




000 


4282.565 


Fc 


5 


4289.885 S 


Cr 


5 


4282.732 




000 


4290.080 


Ti 


I 


4282.860 


Ti 





4290.213 




00 


4282.952 







4290.377 


Ti 


2 


4283.169 s 


Ca 


4 


4290.542 


Fe 


I 


4283.4M 


Ba? 


00 N 


4290.728 




000 


4283.565 




00 


4290.864 




000 


4283.705 




000 


4291.035 




X 


4283.905 




000 


429I.II4 


Ti 


3 


4284.057 




000 N 


4291.174 




ooN 


4284.223 


Mn.V 





4291.276 


Ti 


2 


4284.382 




2Nd? 


4291.375 




I 


4284.565 







4291.630 


Fe 


2 


4284.689 




00 


4291.776 




oooN 


4284.838 


Ni 


I 


4291.896 




ooN 


4284.990 




I 


4291.998 




ooN 


4285.164 


Ti- 


2 


4292.135 


Cr.V 





4285.357 




00 


4292.208 




X 


4285.525 




I 


4292.290 




2 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 297 



Wav«-lei«A 



4292.450 
4292.616 
4292.739 
4292.827 
4292.940 
4293.03s 
4293.192 ) 

4293-273 5 
4293.486 

4293.714 
4293.817 
4293.957 
4294.077 
4294.204 
4294.301 
4294.531 
4294.781 
4294.936 
4295.015 
4295.194 
4295.383 
4295.578 
4295.747 
4295.914 
4296.044 
4296.235 

4296.375 
4296.548 
4296.735 
4296.840 

4296.933 
4297.110 
4297.202 

4297.369 

4297.448 

4297.684 

4297.908 

4298.136 

4298.195 

4298.355 

4298.531 

4298.675 

4298.828 

4298.967 

4299.149 s 

4299.296 

4299.410 

4299.524 

4299.641 

4299.803 



Cr 



Ti 
Fc 

W 
Zr 



Cr.Ti 

Ni 
La 



Zr? 
Cr 



Cr.V 

Ti 
Fc 



Ni 
Ti 

Ca 

Ti,Fc 

Ti 



Inieasity 

•od 
Characttr 



2 

0000 

000 

0000 

0000 

0000 

2 

3 

oooN 

00 

000 

o 

000 

2 

5 

00 

00 Nd? 

2 

0000 

3d? 

3Nd? 

000 

000 

2 

I 

oN 
oN 
oooN 

3 

I 
oN 

2 

X 

2 
2 

iN 
o 

X 

2 
I 

000 
00 

2 
2 

3 

I 

4 
o 
o 

2 



Wavc-lowth 



4299.846 
4299.989 
4300.135 
4300.211 
4300.376 
4300.478 
4300.732 
4300.895 
4300.987 
4301.158 
4301.262 
4301.332 
4301.442 
4301.658 
4301.902 
4302.085 
4302.238 

4302.353 
4302.460 

4302.692 S 

4302.913 
4303.072 
4303247 
4303.337 
4303.584 
4303.754 
4303.880 
4303.992 
4304.098 
4304.300 
4304.415 
4304.552 
4304.729 
4304^82 

4305.013 
4305.266 

4305.377 
4305.479 
4305.614 
4305.772 
4305.871 
4306.006 

4306.078 s 

4306.305 
4306.521 
4306.758 
4306.858 
4307.017 
4307.215 
4307.342 



Ti 
Mn 

Ti 



Ti 



Ti 
Fe 
Ca 



Zr 



Fe,Sr,Ti,Cr 



Ti 



latcntity 

•od 
Character 



00 
xN 



3 
o 
xN 

2 
o 

X 

2 

4 

X 

o 

oN 
oNd? 

2 
00 

2 

2 

4 

2N 

IN 

o 

2 

xN 

I 
o 

2 
4 
00 

X 
X 
2 

o 
ON 

I 
o 

I 

3 
000 

2 

X 

4 

2N 

ooN 
o 

2 
2 

000 Nd? 
000 



298 



HENRY A. ROWLAND 



Wave-letigth 



4307.465 
4307.720 
4307.907 K G 

4308x81 r ^ 

4308.206 

4308.334 
4308.449 
4308.541 

4308.601 

4308.759 
4308.937 
4309.063 
4309.198 
4309.290 
4309.365 
4309.541 
4309.621 
4309.792 
4309.876 

4309.993 
4310.066 
4310.266 
4310.388 
43IO.C40 
4310.631 
4310.722 
4310.862 
43 ".058 
43". 146 
4311.328 
4311.486 
4311.608 

4311.674 
4311.880 
4312.050 
4312.247 
4312.31 1 
4312.462 
4312.665 
4312.723 
4312.871 
4313.034 
4313.193 
4313.399 
4313.577 
4313.797 
4314.052 
4314.248 
4314.381 
4314.479 



Substanoe 



Ca 

Fe 



Fc 
Fc 



Mn 
Ti 



Sc 
Ti 



Intensity 

and 
Character 



2N 

2Nd? 

3 
6 


00 

00 

00 

2Nd? 

000 

I 

2 

o 



3 

I 
I 
I 
o 
o 
2 
I 

2 
I 
O 
2N 

I 

2 

00 

2 

2 

2N 

00 

2 

I 

2 

00 

iN 

o 

3 
iN 

ooN 
00 Nd? 
2Nd? 
00 N 

3 
I 

I 



Wave-length 



4314.673 
4314.894 
4314.964 
4315.138 
4315.262 
4315.446 
4315.626 
4315.769 
4315.901 
4316.032 
4316.115 
4316.246 
4316.720 
4316.832 
4316.962 
4317.125 
4317.226 
4317.484 
4317.618 

4317.883 

4318.064 

4318.232 

4318.369 

4318.522 

4318.631 

4318.817 s 

4318.960 

4319.099 
4319.252 

4319.455 
4319.615 

4319.799 
4319.969 
4320.140 
4320.318 

4320.535 
4320.661 
4320.756 
4320.907 
4321. 119 
4321.294 

4321.395 
4321.583 
4321.678 
4321.813 
4321.961 
4322.204 

4322.373 
4322.521 
4322.670 



Snbitanoe 



Ti 
Ti 
Fe 



Ti? 
Zr 



Ca,Mn? 



Cr 



Sc 



Ti 
Fc 



Cr 



Intenaity 

and 
Character 



ON 

00 

I 

3 

4 

000 N 

ooN 

00 N 

0000 

000 

00 

00 N 

000 N 

000 

I 

o 



oN 

000 

oN 

ooN 

oN 

00 

oNd? 

000 

4 

00 

00 

oooN 

oooN 

o 

o 

oooN 

ooN 

00 

o 



00 

3 

2 

00 

00 

oN 

0000 

o 

2 

oN 

ooN 

00 N 

ON 



TABLE OF SOLAR SPECTRUM WAVELENGTHS 299 







Intensity 






Intensity 


Wave-length 


SlllMtSB^ 


and 
Character 


Wave-length 


Subetanoe 


and 
Character 


4322.862 




OON 


4330.063 




00 


4322.992 




00 N 


4330.189 


V 


ON 


4323.167 




I 


4330.405 




X 


4323.222 







4330.566 







4323.386 




2Nd? 


4330.609 







4323-531 







4330.743 




000 


4323.670 




I 


4330.866 « 


Ti.Ni 


2 


4323772 







4330.984 




00 


4323.8/2 




0000 


4331.II9 






4324.007 




3 


4331.229 




00 


4324.138 




I 


4331.402 




000 Nd? 


4324.246 


Zr 





433X.614 




oN 


4324.340 







4331.81 1 


Ni 


2 


4324.572 




2N 


4331.944 




ooN 


4324.775 




00 


4332.172 




00 N 


4324.886 




00 


4332.339 




00 N 


4324.977 







4332.620 




00 


4325.152 


Sc 


4 


4332.745 




oN 


4325.306 


Ti,Cr 


I 


4332.988 


V 





4325.516 


Ni,Zr 


I 


4333.082 







4325.648 




000 


4333.212 




00 


4325.777 




I 


4333.367 




oN 


4325.939 s 


Fc 


8 


4333.525 




00 


4326.119 




I 


4333.588 




00 


4326.213 







4333.758 




000 


4326.383 




0000 


4333.925 


La 


iN 


4326.520 


Ti 





4334.063 




00 


4326.641 




oooN 


4334.174 




00 


4326.779 




000 N 


4334.228 




00 


4326.923 


Fc 


2 


4334.354 




00 


4327.082 




oNd? 


4334.407 




00 


4327.274 


Fc 


3 


4334.600 




000 Nd? 


4327.319 




00 


4334.833 







4327.479 




oooN 


4334.965 







4327.6X6 




000 Nd? 


4335.XO2 


La 





4327.961 




00 


4335.248 




ooN 


4328.080 


Fc 


2 


4335.434 




iNd? 


4328.202 




ooN 


4335.6x1 




ooN 


4328.365 




000 


4335.762 




00 Nd? 


4328.446 




000 


4335.944 




00 N 


4328.598 




000 


4^36.076 




00 


4328.772 




oN 


4336.X47 




000 


4329.007 




00 


4336.295 




0000 


4329.089 




000 


4336.433 




00 


4329.213 




00 


4336.602 




000 Nd? 


4329.306 




0000 


4336.774 




00 N 


4329.448 




oN 


4336.951 




00 


4329.559 




oN 


4337.034 




00 


4329.712 




00 


4337.2x6 


Fc 


5 


4329.853 




00 


4337.414 








'This is a weak, haiy Ni line. It is faintly present in a specimen of meteoric 
iron. The Ni line is on the red edge of the solar line and the Ti line is nearer the 
center. 



300 



HENRY A. ROWLAND 



Wave-length 



Sttbetaiioe 



4337.569 
4337.725 
4337 945 
4338.084 
4338.245 
4338.29a 
4338.430 
4338.616 
4338.779 
4338.854 
4338.993 
4339.170 
4339 294 
4339.416 

4339.617 

4339.882 

4340.070 

4340.192 

4340.297 

4340.634/^ 

4341.007 

4341.167 

4341.286 

4341.410 

4341.530 

4341.723 

4341.880 

434X.991 

4342.087 

4342.219 

4342.350 

4342.482 

4342.750 

4343-004 

4343.150 ^ 

4343-372 ) s 

4343.431 S 

4343.577 
4343.662 
4343.861 
4344.015 
4344.130 
4344.306 
4344.451 
4344.670 
4344.830 
4344.906 
4345.052 
4345.246 
4345.399 



Mn 
Cr 
Sr? 
Ti 



Fe 



Fe 
Cr 
Cr 



Cr 
H 

V 
Zr 

Ti? 



Cr 
Fe 



Fe 



Ti- 
er 



Intensity 

tnd 
Character 



00 

3 
00 

4 

000 

00 

I 

od? 

00 

o 

oN 

000 

o 



4 

3 

000 

000 

o 

20 N 

00 Nd? 



0000 

00 

2 

00 





o 

000 

o 

00 

oooN 

000 N 
ooN 
2 

2 

ooN 

oN 

2 

0000 

1 N 
ooN 
2 

4 

000 

00 



00 

0000 



WaTe*lei«di 



4345.506 
4345.593 
4345.767 
4345.933 
4346.068 
4346.278 

4346.454 
4346.580 
4346.725 
4346.835 
4346.987 
4347.066 

4347.199 
4347.266 
4347.403 
4347.532 
4347.705 
4347.845 
4348.003 
4348.130 
4348.266 
4348.349 
4348.497 
4348.651 
4348.800 
4348.933 
4349.107 
4349.330 
4349.538 
4349.967 
4350.U9 
4350.319 
4350.413 
4350.550 
4350.747 
4350.921 
4351.000 
435x216 
4351.338 
435x464 
4351.551 
4351.71 1 
4351.871 
435X930 
4352.083 
4352.235 
4352.425 
4352.559 
4352.7x8 

4352.908 S 



latentlty 





0000 








ooN 




00 




oN 




00 




t 




00 


Fe 


2 




00 


Cr 


t 




0000 




000 




000 


Fe 


I 




ooN 




iN 




000 


Fe 


2 




iN 




000 




00 




IN 




ooN 




ooN 




ooN 


Fe 


2 




oooN 




000 




00 




ooNd? 












Be? 


ooN 


Fe 


ON 






Ti 


t 


Cr 


3 




00 




00 




00 


Fe 


3 




0000 


Cr 


5 


Mg 


SNd? 




ON 




oN 




ooN 




oN 


Fe 


4 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 30I 







iBtentily 






Intanaily 


Wm-lcBcA 


Sabttanoe 


umI 


Wave-leagth 


Subatance 


and 






Chancier 






Characier 


4353.044 


V 





4360.451 




I 


4353.214 




OOON 


4360.644 


Ti 


I 


4353.337 




OON 


4360.796 




000 


4353.596 




ooN 


4360.958 


Fe.Zr 


I 


4353.678 




ooN 


4361.091 




000 


4353.813 




ooN 


4361.219 




000 Nd? 


4353.985 




00 


436MO5 




00 


4354.110 







4361.474 




00 


4354.228 




00 


4361.676 




0000 


4354.426 




oNd? 


4361.828 




000 


4354.597 




00 


4361.955 


Sr? 


ooN 


4354.674 




00 


4362.021 




0000 


4354.776 




I 


4362.198 




000 


4354-927 


Fe? 


oN 


4362.262 







4355.112 




ooN 


4362.387 




00 


4355.257 


Ca? 


2 


4362.542 




000 


4355.509 







4362.690 




I 


4355.578 




00 


4362.909 




ON 


4355.746 




00 


4363.113 




ooN 


4355.867 







4363.267 


Cr 


IN 


4356.058 







4363.457 




oN 


4356.163 


Ni 





4363.629 







4356.299 




00 


4363.766 







4356.414 




000 


4363.979 




ooN 


4356.528 







4364.136 




000 


4356.766 







4364.198 




I 


4356.901 




00 


4364.349 




I 


4357.071 







4364.490 




00 


4357.223 




000 


4364 662 




ooN 


4357.312 




ooN 


4364.827 


Ce 


00 


4357.457 




00 


4365.031 




000 


4357.677 




oNd? 


4365.168 




000 


4357.865 




000 


4365.451 




ooN 


4358.033 







4365.694 







4358.178 




00 


4365.885 




00 Nd? 


4358.328 







4366.061 


Fe 


2 


4358.521 




00 


4366.246 




ooN 


4358.670 


Fe 


2 


4366.362 




ooN 


4358.879 


Y-Zr 





4366,572 




00 


4358.985 




00 


4366.660 


Zr- 


I 


4359.084 




00 


4366.838 




I 


4359.236 




000 


4367068 




00 


4359.358 




00 


4367.222 




000 


4359.500 




00 


4367.354 




00 


4359.654 


Ni 





4367.495 




00 


4359.784 s 


Cr 


3 


4367.637 




00 


4359.907 


Zr 





4367.749 


Fe 


5 


43604)67 




00 


4367.839 


Ti 


2 


4360.150 




00 


4367.882 




00 


4360.280 




0000 


4368.071 


Fe 


2 



302 



HENRY A. ROWLAND 



WaYe-lcngth 



4368.224 
4368.291 
4368.462 
4368.629 
4368.801 
4369.052 

4369.254 
4369.428 
4369.568 
4369.708 

4369.873 
4369.941 S 

4370.025 
4370.x 89 
4370.318 
4370.453 
4370.583 
4370.743 
4370.815 
4371.016 

4371.144 
4371.221 
4371.320 
4371.442 
4371.588 
4371.748 
4371.957 
4372.1x5 
4372.184 
4372.360 
4372.498 
4372.656 
4372.747 
4372.899 
4373.008 
4373.148 
4373.278 
4373.415 
4373.554 
4373.727 
4373.813 
4373.949 
4374.055 
4374.206 

4374.331 
4374.384 
4374.628 

4374.775 
4374.981 
4375.103 



Ni 
Ti 



Ti 
Fc 

Ni 



Zr 
Cr 



Fc? 
Cr 
Fc 
V 

Cr 

Sc, Fc? 

Zr 
V. Mn 



Intensity 

and 
Character 



00 

O 

O 

000 



00 Nd? 

00 

00 

I 

ooNd? 



4 

00 

00 

o 

o 

00 

000 

00 

000 

X 
X 
00 

2 

oNd? 

000 

oN 

00 

00 

000 

od? 

00 

0000 



o 


000 

I 

ooN 

2 

0000 





000 

X 

I 

3 

00 N 
o 

2 



Wave-length 



4375.216 

4375.358 

4375-493 

4375643 

4375.734 

4375.823 

4375.940 

4376.107 s 

4376.38 X 

4376.572 

4376.727 

4376.942 

4377.X13 

4377.254 

4377.388 

4377.533 

4377.706 

4377.948 

4378.149 

4378.419 

4378.677 

4378.902 

4379.071 

4379.232 

4379.396" 

4379.565 

4379.691 

4379.798 

4379.927 

4380.066 

4380.325 

4380.395 
4380.521 
4380.655 
4380.883 
4381.004 
4381.146 
4381.274 
438X.451 

4381.875 
4382.045 

4382.159 
4382.325 

4382.475 
4382.676 

4382.847 
4382.928 
4383.156 
4383.332 
4383.468 



Cr 



Fc 



Fc.Cr 



Fc 



Zr 
Co 



Cr 
Mn 



Mn 
Fc 



Intensity 

and 
Character 



000 

oNd? 

00 


o 

0000 

6 

ooN 

ON 

oN 

X 

ooN 
ooN 

2N 

od? 

o 

I 

ooN 

2Nd? 

oNd? 

ooN 

iN 

ooN 

4 

000 

00 

000 



000 

2Nd? 

000 

00 



2Nd? 

ooNd? 

00 

o 

ooN 

00 

00 

00 

00 

ooN 

o 

o 

2 

oN 
fooN 
oooN 



' The strongest Vanadium lines in the whole solar spectrum or in this part, if not 
the whole, of the Vanadium spectrum. 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 303 



WaTe-leugth 



4383.535 

4383.720 S 

4383.879 

4383.990 

4384.122 

4384.284 

4384.477 

4384.698 

4384.873' 

4384.986 

4385.144 
4385.286 
4385.406 
4385.548 
4385.767 
4385.833 
4386.016 
4386.221 

4386.433 

4386.616 

4386.750 

4386^55 

4387.007 

4387.220 

4387.420 

4387.559^ 

4387.658 \ 

4387.762 J 

4387.906 

4388.057 

4388.260 

4388.411 

4388.571 

4388.746 

4388.887 

4389.029 

4389.190 

4389.247 

4389.413 

4389.555 

4389.667 

4389.801 

4389.930 

4390.034 

4390.149* 

4390.273 

4390.379 
4390.496 
4390.617 
4390.699 



Fe 



Ni 
V 

Cr 

La 



Ti? 



Cr 



Fc, Co 
Md 

Fc 



Fc- 

Mn, Ni? 
V 

Ni 
Fc? 



Intensitj 



Character 



000 N 

15 

000 N 
000 N 
oooN 
ON 
I 

3 



2 

I 

1 

2 

00 

00 

ooN 

ON 

0000 Nd? 
o 

00 

ooN 

I 

1 N 
oooN 


o 

o 

000 

2 

00 Nd? 

00 N 

3 

ooN 

o 

oN 

000 

00 

2 

000 

oN 





00 

2 

000 

00 

00 

I 

o 



Wave- length 



4390.790 

4390.932 

4391.002 

4391.123 { , 

4391.192 J • 

439^306 

4391.462 

4391.642 

4391.824 

4391.924 

4392.034 

4392.235 

4392.470 

4392.591 

4392.752 

4392.947 

4393.085 

4393.196 

4393.442 

4393.686 

4393.858 

4393.974 

4394.093 

4394.225 

4394.342 

4394.463 

4394.782 

4394.943 
4395.020 
4395.201 

4395.413' 

4395.448 

4395.665 

4395.843 
4396.008 
4396.224 
4396.312 
4396.471 
4396.592 
4396.790 
4396.931 
4397.X25 
4397.306 
4397.428 
4397.550 
4397.742 
4397.952 
4398.178 

4398.334 
4398.460 



Suhttanoe 



Fc 
Ti 



Co 
Cr 
Co 
V? 



V? 



V? 
Ti 
Ti? 



Zr 

Ti 

V, Zr 



Ti 



Intensity 

and 
Character 





00 N 

0000 

2 

I 

000 

00 

00 



I 

o 

iN 

00 N 

000 N 

I 

ooN 

00 

o 

o 

iNd? 

o 



o 

2 

00 

000 

000 

00 

00 

3 

2 

00 



oooN 

I 

00 
00 

o 

00 

00 N 

000 

iN 

00 
00 

00 

000 N 
00 Nd? 

I 

00 





' The strongest Vanadium lines in the whole solar spectrum or in this part, if not 
the whole, of the Vanadium spectrum. 

'Very strong Vanadium lines. sjn Zircon but not in Zr. 



304 



HENRY A. ROWLAND 



Wave-length 



4398.651 
4398.784 
4398.879 
4399.010 
4399.230 
4399.379 
4399.451 
4399.640 
4399.776 
4399.935 
4400.149 
4400.343 
4400.555 
4400.738 
4400.844 
440 X. 020 
4401.183 
4401.241 
44OX.456 
4401.613 
4401.709 
4401.828 
4402.509 
4402.640 
4402.843 
4403.111 
4403.241 
4403.347 
4403.532 
4403.661 

4403.815 
4403.990 
4404.131 
4404.261 

4404.433 
4404.563 
4404.706 

4404.757 

4404.927 s 

4405.082 

4405.191 

4405.472 

4405.576 

4405.728 

4405.896 

4406.201 

4406.319 
4406.456 
4406.662 
4406.8x0' 



Ni 

Ni 
Ti,Cr 

Zr 

Sc 
V 

Ni 



Fe 
Fe 

Ni 



Cr.Zr 
Cr 



Ti 



Fc 



loieniitj 
and 

Character 



ON 

00 

00 

00 

ooN 

00 



00 



3 

ooN 

oNd? 

3 
I 

000 
oN 
iN 
000 
2 
I 
2 

000 
ooN 
ooN 
oooN 
00 N 
00 
I 
o 
00 
ooN 
ooN 

00 
iN 
00 

fooN 
ooN 

Loo 
I 

000 
00 
ooN 
oNd? 
00 


000 
00 
2 



^^avc'leiigth 



4406.997 
4407.155 
4407.297 
4407.432 
4407.533 
4407.680 
4407.810 ) 

4407.871 5 

4408.022 

4408.096 

4408.236 

4408.364" 

4408.582 

4408.683* 

4408.818 

4408.956 

4409.096 

4409.288 

4409.408 

4409.525 

4409.683 

4409.853 

4410.015 

4410.168 

4410.328 

4410.463 

4410.683 

4410.817 

4410.923 

4411.019 

44x1.111 

4411.240 

4411.385 

4411.748 

4411.882 

44«.043 
4412.092 

44x2.297 
4412.415 
4412.581 
4412.861 
4413.042 
4413.280 
4413.560 
4413.756 
4413.944 
4414.OII 
4414.207 
4414.278 
4414.391 



lotenstty 
and 

Character 



V 
Fe 



V 
Fc 
V 



Fc? 



Cr 

Ni 



Cr- 



Mn 

V 
Cr 



Zr 



Cr 



00 

00 

000 

oN 

ooN 

000 

2 

4 

000 

00 

000 

2 

3 

2 

00 

ooN 

ooN 

I 

o 

00 N 

I 

oooN 

oooN 

ON 

ooN 

DON 

2 

000 

0000 

00 

000 

I 

ooN 

ooN 

ooN 

000 

I 

00 



ooN 

ooN 

ooN 

ooN 

ooN 

I 

000 

o 

000 

00 

00 



'Very strong Vanadium lines. 



PLATE XIV 



*IOO -75 -50 -28 O tf 10 ft 109 



t 








e 
















/ 


\ 








10 








/ 


\ 












/ 






V 






11 




c 


/ 






\ 


• 






J 










v 




Ift 




r 










\ 




/ 














\ 





















L 



100 -75 -SO -as o %% to 1% too 



Fig. I 





Fig. 2 



Fig. 3 



T ANDROMEDiE. 

By Edward C. Pickering. 

On learning of the discovery by Mr. Anderson of the variable 
star T Andromedae, an examination was made of the Henry 
Draper Memorial photographs of this object. The results 
were communicated to the Astronomis^he Nacknchten (1349 347), 
and, as there stated, indicate a photographic magnitude of 9.0 
at maximum, and a very uniform increase and diminution in the 
light at the rate of one magnitude in twenty-six and twenty-five 
days respectively during the three months preceding and follow- 
ing the maxima. This form of light curve is confirmed by the 
photographs taken since then which are enumerated below. 
They also indicate a change in the period, the value 281 days, 
which satisfies the observations during 1891 to 1894, giving a 
maximum later than that which actually occurred in 1895. 
These results are represented in the following table, which gives 
in successive lines all the photographs so far obtained here of 
this star. Photographs of the region taken when the star was 
too faint to appear are not included. The dates and observed 
photographic magnitudes are followed by the maximum com- 
puted by the law given above. Thus the first plate was taken 
on the Julian Day 2,412,039. The niagnitude 10.4 indicates 
that it was 1.4 magnitudes fainter than the maximum, and multi- 
plying 1 .4 by 26 gives 36, the time in which the maximum would 
be attained. Adding 36 to 2,412,039 gives 2,412,075, or 2,075 
if we omit the constant 2,410,000. This quantity is entered in 
the third column. The mean of the individual values of each 
time of maximum is given in the next column, followed by the 
residuals found by subtracting it from the individual values. 
The residuals have the average value of ± 2,9 days, correspond- 
ing to a deviation of the observed magnitudes of ± o.i i. These 
values would be reduced one-quarter if we could reject the last 
three results. The latter are not due to errors of observation, 

305 



3o6 



EDWARD C. PICKERING 



since a second independent measurement gave the same result in 
each case within 0.05 of a magnitude, corresponding to a change 
of two days in the time of maximum. While these photographs 
fail to show whether the light curve is pointed or rounded at 
the exact time of maximum, they indicate that the curvature, if 
any, is inappreciable except within a few days of the maximum. 



Date 


Ob«.M«R. 


Mm. 


Me«B 


RMid. 


1891 Nov. 2 


10. 4 


2075 


2074 


4- X 


" " 27 


9.3 


2072 




— 2 


" Dec. 13 


9.2 


2075 




-f X 


1892 Oct. 24 


10. 6 


2356 


2356 




" Nov. 6 


II. I 


2357 




+ ? 


" " 13 


n.4 


2356 






1893 Sept. 13 


".3 


2638 


2638 




1894 Jan. 2 


".3 


2917 


29X6 


+ I 


" " 7 


12.0 


2914 




-^ 2 


« •! ^ 


12.0 


2914 




— 2 


.1 « ,^ 


".5 


29x3 




— 3 


« « 25 


".5 


2919 




+ 3 


" Feb. 2 


II. I 


2917 
3168 




+ I 


•• Sept, 28 


II. 6 


3x73 




" Oct. M 


II.4 


3175 




x\ 


« " 18 


II. I 


3175 




•* " 19 


no 


3173 






" " 20 


10.8 


3169 




— 4 


•• Nov. 6 


10.2 


3x70 




-« 3 


" Dec. 5 


9.0 


3168 




"^ 5 


•• ** ^ 


8.8 


3169 




~" 4 


" " 13 


9.0 


3x76 




+ 3 


•• " 30 


9.6 


31M 




+ 1 


X 895 Jan. 2 


9,6 


3181 




" Feb. 5 


10.8 


3185 




+ u 



The form of light curve is shown in Plate XIV, Fig. i, in which 
abscissas represent the times in days preceding or following the 
observed maxima, and ordinates the corresponding magnitudes. 
The assumed law is represented by the heavy line. The obser- 
vations from which the form of light curve was inferred are rep- 
resented by crosses, the later observations by circles. 

If we take first differences of the times of maxima found 
above we find the intervals 282, 282, 278 and 257. The obser- 
vations during 1 89 1 to 1894 are therefore very well satisfied by 
the period 281 days. The observations of 1894 indicate a change 



T ANDROMEDjE 307 

in the period which cannot be accounted for by errors of obser- 
vation. Rejecting the last two observations changes the mean 
time of maximum from 3173 to 3 171. 

The magnitudes given above are found by comparing the 
photographic images of the variable with adjacent comparison 
stars of nearly equal brightness and estimating the difference in 
magnitude. The comparison stars when the variable is bright 
are BD. + 26^ 40, + 26° 47i + 25"" 42 and + 25° 40. Their 
photographic magnitudes are 8.7, 9.2, 9.8 and 10.3. The posi- 
tions and magnitudes of the fainter stars will be given elsewhere. 

In a recent article in the Astranmmcal Journal (14, 183) Mr. 
P. S. Yendell describes his observations of this star and con- 
cludes that the period is 265.35 days. Since the photographic 
magnitudes do not accord with this theory, he derives the sin- 
gular conclusion that they must be wrong, maintaining that some 
of them are in error by two or three magnitudes. He states that 
the light curve described above has a form which **is not only 
inherently improbable, but which actually proves to be incorrect." 
Inherently improbable does not seem to be a strong argument 
in view of the variety in form of light curves of variable stars, 
especially as the linear form is strikingly confirmed by Nova 
Aurigae and other variables {A. N, 1341 138). It also represents 
one of the simplest theoretical laws, the variation in the energy 
being proportional to the energy itself. Furthermore it coin- 
cides with Newton's law of cooling. Whether the light curve is 
actually incorrect cannot be proved by observations made at a 
different time and on a different portion of the light of the star. 
As Mr. Yendell does not give the light curve he has himself 
deduced or even the names and magnitudes of his comparison 
stars, it is difficult to discuss it. It may be noted, however, that 
unless he measured the variable photometrically or used magni- 
tudes photometrically determined for his comparison stars, it 
would be impossible for him to tell whether his light curve was 
linear or not, when represented on the scale of Pogson used 
here. The error in his assumed scale of magnitudes might 
easily introduce a marked deviation from a straight line in his 



308 EDIVARD C. PICKERING 

curve. Mr. Yendell's table of observed maxima appears to me 
illusory. Two of these maxima are derived from the Harvard 
photographic magnitudes by a process which he does not 
describe, and give results which differ widely from those found 
here. It is surely impossible to infer from maxima thus obtained 
that the observations on which they depend are themselves 
sometimes in error by more than two magnitudes. If such 
errors exist, this determination of the maxima should have been 
rejected. 

The photograph of this star taken on November 27, 1891, is 
represented in Fig. 2. The variable, A, is distinctly brighter 
than the two stars below and to the right, which are of about the 
tenth magnitude. The upper of these stars is BD. + 25** 40. 
A defect in the original negative to the left of this star has been 
removed in the print without affecting the image of the star 
itself. The magnitude of the variable on this day according to 
the above table is 9.3. The photograph taken October 24, 1892, 
is represented in Fig. 3. The variable, B, is here fainter than 
the stars below. Its magnitude is given above as 10.6, or 1.3 
magnitudes fainter than A. According to Mr. Yendell's theory 
on the first date the photographic magnitude 9.3 is too bright 
by 2.1 magnitudes, and its magnitude should have been 11.4. 
B similarly should have been 10.5, or 0.9 magnitudes brighter 
than A. An inspection of the plate will enable the reader to 
decide whether to believe that Mr. Yendell's theory or the photo- 
graphic magnitudes are in error. 

Harvard College Observatory, 

Cambrwge, Mass., 

March 5, 1895. 



ECLIPSE OF JUPITER'S FOURTH SATELLITE. FEB- 
RUARY 19, 1895. 

By Edward C. Pickering. 

Photometric observations of the satellites of Jupiter while 
undergoing eclipse have now been maintained at the Harvard 
College Observatory for many years. This method has special 
advantages in the case of the fourth satellite, owing to the slow 
variation in its light. The total number of eclipses of this satel- 
lite that can be observed from any one station is comparatively 
small. From April 16, 1892, to January 16, 1895, ^^^ satellite 
passed outside of the shadow of Jupiter and was not eclipsed. 
On January 16 Jupiter was below the horizon at Cambridge at 
the time of the eclipse. At the next eclipse on February 2 the 
night was cloudy. 

On February 19, 1895, ^^^ fourth satellite was compared 
with the second satellite by means of a photometer consisting of 
a double image prism and Nicol {Harvard AnnaJs, zi| 4). Obser- 
vations were made continuously from 28 minutes before to 26" 
42* after the predicted time of disappearance, 13** 23" i8*.4 
Greenwich Mean Time. Observations of the reappearance, 
which was predicted to occur at 14** 58" 52*.o G. M. T., were 
prevented by clouds. The photometer was attached to the 
15-inch telescope of the Observatory, and all of the observations 
were made by Mr. O. C. Wendell. In the annexed table the 
observations are arranged in groups in successive lines. Each 
group consists of twenty photometric settings, except the last, 
which consists of eight settings. The mean of the observed 
times minus the computed time of disappearance is given in the 
first column expressed in seconds, and the corresponding mean 
difference in magnitude of the fourth and second satellites in 
the second column. As the satellites are close together they 
would be equally affected by the clouds which interrupted the 
later observations. The deviations of the observed magnitudes 

309 



310 



EDWARD C. PICKERING 



from a smooth curve, whose second diSFerences are nearly con- 
stant, are given in the third column. The average value of these 
differences is only ±.04 so that the form of the curve is defined 
within narrow limits. 



Time 


Ob«.M«R. 


o.-c 


rime 


OlM.MaK. 


0.-C 


~I564* 


X.OO 


+.05 


-i- 202* 


2.20 


-f.05 


— X292 


.94 


—.03 


+ 395 


2.38 


—.04 


- 850 


1. 16 


+.03 


-f- 606 


2.74 


— .01 


— 6X2 


1.24 


—.04 


-f 802 


3.10 


+.02 


- 389 


1.36 


—.11 


+ 1054 


3.56 


+ .01 


— X94 


X.64 


— .02 


+ 1508 


4.42 


— .01 


+ 3 


1.96 


+ .07 









The large deviation of the time of disappearance from that 
given by computation, and the long duration of the variation are 
in part due to the obliquity with which the satellite entered the 
shadow of Jupiter. 

Harvard College Observatory, 

Cambridge, Mass., 

March 5, 1895. 



THE SPECTRUM OF MARS. 
By Lewis E. Jewell. 

There has been a discussion in recent numbers of Astronomy 
and Astro-Physics between Professor Campbell and Dr. Hug- 
gins on the spectrum of Mars. As I have recently been mak- 
ing a careful spectroscopic study of water-vapor in the Earth's 
atmosphere, I have thought that some of the results bearing upon 
the possibility of determining the presence of water-vapor in the 
spectrum of Mars might be of interest. 

In the investigation just referred to the quantitive values of 
the intensities of the lines in the spectrum of water-vapor were 
determined in terms of the amount of water-vapor the sun- 
light was required to pass through in order that the water-vapor 
lines might be of any given intensity. These values were deter- 
mined by an indirect method, so that the absolute values may be 
a little too large or too small, but the relative values are correct, 
and the absolute values are approximately correct. I have also 
determined the amount of water-vapor required to make its pres- 
ence visible in the spectroscope. For this purpose several spec- 
tros/:opes have been used, but the results will be given for only 
three of these instruments. The large concave gating of 20,000 
lines to the inch and twenty-one feet six inches radius gives an 
actual separation of the D lines of 6"*". 14 at the focus of the 
grating in the second spectrum without any magnification. 

A plane grating of 7200 lines to the inch, used with col- 
limating lenses of six and one-half inches diameter and seven 
feet six inches focal length, gives a separation in the first spectrum 
of o™.826. 

A Steinheil spectroscope of two 6o^prisms, used with coUimat- 
tng lenses of one and one-half inches diameter and thirteen inches 
focal length, gives a separation of 0°*°*.056. This spectroscope 
shows the nickel line between the D lines, but it is difficult to 
see. These spectroscopes used in the manner indicated will be 

3x1 



3 1 2 LEWIS £. JEIVELL 

referred to as the concave grating, the plane grating, and the 
Steinheil spectroscope. 

With the concave grating an amount of water-vapor equiva- 
lent to a thickness of 0.23 inches of water can be detected ; but 
the amount cannot be satisfactorily measured until it is equiva- 
lent to one-half or three-quarters of an inch of water. The 
accuracy of determinations made with this grating is largely due 
to the fact that the individual lines in the spectrum of water- 
vapor are distinctly visible, and the dispersion is so great that an 
extremely faint and narrow line can be detected. With the plane 
grating ( used for this purpose only in the first spectrum ) the 
dispersion was much less, and it was impossible to detect a line 
unless it was three or four times as strong as the faintest lines 
seen with the concave grating. By observing the intensity of 
the strongest line in what is known as the "rain-band" (wave- 
length =59 19.8 5 5 ) an amount of water-vapor equivalent to 0.67 
inches of water can be detected ; but it would require as much as 
one and a half or two inches to enable one to make reasonably 
accurate observations. 

If instead of using this single line we observe the group of 
three prominent lines of which this line forms the red component, 
we are but little better off, as the violet component of this trip- 
let is partly solar. This triplet is, however, the best group of 
lines for this purpose in the spectrum of water-vapor. 

With the Steinheil spectroscope none of the individual lines 
in the spectrum of water-vapor can be seen ; the triplet referred 
to can, however, be seen as a single line, with an amount of 
water-vapor equivalent to one and a half inches of water. It 
was, however, seen with the greatest difficulty ; and no reason- 
ably good observation can be made of the intensity of this triplet 
unless the sunlight has passed through an equivalent of three or 
four inches of water; in fact, considerably more would be 
required to make satisfactory estimates. If instead of observing 
this triplet we observe the group of lines between the wave- 
lengths 5914 and 5925, which Professor Campbell has called d" 
we will have but little better results than with the triplet. The 



THE SPECTRUM OF MARS 313 

region //* contains both water-vapor and solar lines, and with a 
dispersion so small as to necessitate the use of this group, the 
presence of water-vapor cannot certainly be detected until the 
group of water-vapor lines is about half as intense as the group 
of solar lines. This group is without doubt the best one for the 
purpose of any of those used by Professor Campbell. 

The values given are approximately correct when the Sun is 
used as the source of light and the definition is very good. Where 
a weak source of light is used, such as the Moon or the planet 
Mars, the quantity of water-vapor required to make its presence 
visible in the spectroscope, may be considerably greater than the 
values given. 

We will now consider the amount of water-vapor present in 
the air during the different seasons of our own planet. We will 
express it by the depth of a layer of water, equivalent to the 
amount of water-vapor contained in the air. The amounts given 
in the following table are the monthly means as observed at 
Baltimore. I also give the amounts present upon two special 
occasions, viz,^ January 10, 1893, and the mean of July, 14, 15, 
16, 17 and 18, 1893. 

January o*".73 August 2*".i7 

February o .86 September i .56 

March o .95 October i .46 

April I .28 November i .04 

May 2 .17 December o .77 

June 3 .25 January 10, 1893 o .31 

July 2 .45 July 1 4. 1 5» 1 6, 17, 1 8, 1 893 (mean) 5 .40 

It is perhaps well to repeat that the quantities as given in the 
table are only approximate, being determined by an indirect 
method, and the true values may be somewhat greater or less 
than the values given, but they are probably not far from right ; 
and the relative amounts are fairly well determined. 

To better understand what we should expect to see in the 
spectrum of Mars, let us imagine what we should see were we to 
examine the spectrum of the Earth's reflected light, from a 
suitable point in space. Let us suppose we are examining the 
Earth's spectrum with a spectroscope whose slit stretches all the 



3M 



LEWIS E. JEWELL 



way across the image of the Earth as formed at the focus of the 
telescope we are using. Also, to avoid complications, let us 
further suppose the atmospheric conditions to be uniform over 
the Earth at the time, irrespe;ctive of latitude ; and let us con- 
sider the Sun to be in the zenith of the place at the center of the 
image of the Earth. 

If we suppose the atmospheric conditions to be similar to 
those prevailing during January (as observed at Baltimore) the 




The intensity of a water-vapor line in the spectrum of the Earth's atmosphere as 
seen from a point in space. 
I, Curve for January. 
II, " •' October. 
Ill, " " June. 

£ and W the edges and C the center of the image of the Earth as seen in a tele- 
scope. 

intensity of the water-vapor lines will vary from the center to the 
edges of the spectrum, as in Curve I in the accompanying figure. 
The conditions prevailing during October and June are repre- 
sented by Curves II and III. 

However, to better show the difficulties of such observations, 
the curves do not exactly represent the relative intensities of the 
lines in the spectrum of different portions of the disk, but repre- 



THE SPECTRUM OF MARS 3 ' 5 

sent the amount of water-vapor the sunlight passes through in 
twice traversing the air. The scale at the side gives the amount 
in inches of water. The slope of curves representing relative 
intensities would be somewhat less steep than the curves given ; 
otherwise they would be similar. 

We could detect the presence of water-vapor in the spectrum 
of the Earth's atmosphere if the conditions were similar to those 
prevailing during January in the latitude of Baltimore, providing 
the dispersion and definition were equal to that of the plane 
grating when using the Sun as the source of light. With a dis- 
persion and definition equal to that of the Steinheil spectroscope, 
we could detect the presence of water- vapor in the Earth's 
atmosphere if the hygroscopic conditions were similar to those 
prevailing at Baltimore during October. 

The spectrum of Mars and the Moon, as seen by Professor 
Campbell^ seems to have been inferior in point of definition to 
that given by the Steinheil spectroscope used with sunlight, and 
the instrumental equipments of Dr. Huggins and the other 
observers who have worked upon the spectrum of Mars were 
probably much inferior to those of the Lick Observatory. As a 
consequence, it necessarily follows that unless the amount of 
water in the atmosphere of Mars is greater than that in the 
Earth's atmosphere in October at Baltimore, it is useless to look 
for the presence of water-vapor in the spectrum of Mars, unless 
our instrumental means are much superior to any hitherto used 
for that purpose. The subject is rendered much more difficult 
by the fact that when we examine the spectrum of Mars we are 
looking through a large amount of water- vapor in the Earth's 
atmosphere. 

Professor Campbell strenuously objects to other observers 
making observations upon Mars when the planet's altitude is 
small and the humidity of the air considerable ; and yet his own 
observations were made during those months when the air con- 
tains the greatest amount of water-vapor. He also seems to 
consider the relative humidity the chief factor to be considered, 
whereas it is the dew-point that is important. We may have a 



3 1 6 LEWIS £. JEWELL 

low relative humidity during warm weather and still have a very 
large amount of water-vapor in the air ; and we may also have a 
high relative humidity during cold weather and have very little 
water-vapor in the air. It is also well to say that the surface con- 
ditions as determined by the hygrometer are not necessarily the 
index of the amount of water-vapor present in the air above the 
surface of the Earth. At times the discrepancy is considerable. 
Neither is he entirely right in his estimate of the great advan- 
tages resulting from an altitude of 4200 feet in observations upon 
the spectrum of the planets. There is unquestionably an advan- 
tage, but it is much less than he thinks, especially during the 
months in which he made his observations. The distribution 
of water-vapor in the air is not similar to the distribution 
of oxygen and nitrogen. During very cold weather and in 
a high barometer there is some similarity, but during the 
warm, humid months the amount of water-vapor in the air 
increases with the altitude to near the height of the lower clouds, 
and then begins to decrease. As a consequence there is less 
advantage in a moderate elevation than we would expect. During 
very cold weather with a high barometer, however, the advan- 
tages are greater. 

Unless the aqueous contents of the atmosphere of Mars 
are much greater than we are justified in assuming from the 
appearance of the planet, it seems useless to attack this 
problem with such spectroscopes as have hitherto been used ; and 
instruments of greater dispersion are unsuitable because of the 
lack of sufficient light. 

In regard to determining the presence of oxygen, however, 
the case is not quite so hopeless, if we make our observations at 
a considerable elevation with the planet near the zenith, for the 
B group can be readily seen with small dispersion, and the pres- 
ence of oxygen in the atmosphere of Mars might possibly be 
detected by careful observations. 

It seems to me, however, that there may be more hope of 
detecting the presence of chlorophyl than either oxygen or water 
vapor in the spectrum of Mars. Of course it will require careful 



THE SPECTRUM O? MARS 317 

and delicate observations, but the chlorophyl band in the red end 
of the spectrum of vegetation is quite strong ; and if the green 
areas of Mars are due to vegetation, this band might be seen. 

In speaking about this matter with Professor W. H. Pickering 
a few weeks ago, I asked him why no one had looked for this 
band, and he informed me that he had looked for it, but that 
the season on Mars was so far advanced at the time that the 
areas that had been of a strong green color had become gray at 
the time his observations were made, so that they were not con- 
clusive. 

If the green areas upon Mars are due to vegetation, and the 
ruddy areas similar to the desert regions of the Earth, the chlo- 
ryphyl band ought to be recognized by its presence in the spec- 
trum of one area and its absence in the other. 
Johns Hopkins University. 



ON A NEW METHOD OF MAPPING THE SOLAR 
CORONA WITHOUT AN ECLIPSE. 

By George £. Hale. 

In the October (1894) number of Astronomy and Astro- 
Physics I described the various methods that have been employed 
for the purpose of observing the corona without an eclipse. 
Unfortunately, I was unable to show that any measure of success 
had attended the numerous investigations conducted with this 
end in view, nor was it possible to conclude the paper with the 
expression of any very confident hope of an early solution of the 
problem. 

An account of Professor Riccb's experiments with the spec* 
troheliograph which I had left on Mount Etna was published in 
the January number of The Astrophysical Journal. In spite 
of the fact that the atmospheric conditions enjoyed by Professor 
Ricc6 were certainly excellent, the method failed of success. 
The coronal forms shown on the best negatives obtained are 
probably of atmospheric origin. They fail to show any of 
the true coronal structure, and the intensity of the halo around 
the solar image falls off gradually and uniformly as the distance 
from the limb increases. On account of the brightness of the 
sky it was necessary to give an exposure much too short for the 
comparatively feeble light of the corona. It is true that the 
speculum had been considerably tarnished by the fumes of the 
volcano, and with a perfect mirror the photographs would have 
been better. But a careful study of the experiments and results 
has led to the conclusion that even with perfect apparatus it 
would have been practically impossible to photograph the corona 
by the method employed, without much superior and possibly 
unattainable atmospheric conditions. 

It would thus seem that the K band offers insufficient pro- 
tection against the brilliant light of the sky. The photo- 
graphic plate is unable, even when exposed only to this com- 

318 



METHOD OF MAPPING THE SOLAR CORONA 319 

paratively dark band, to distinguish between the light of the sky 
and that of the corona and sky together. During an eclipse 
the corona is photographed with the greatest ease, but when the 
bright light of the sky and the feeble coronal image meet the 
plate simultaneously the photographic method breaks down. On 
account of the inability of the eye or the sensitive plate to 
detect very small di£Eerences in the brightness of various parts 
of an illuminated object, the methods proposed for observing or 
photographing the corona in sunlight have almost without excep- 
tion been based upon some plan of increasing the contrast 
between the corona plus sky and the sky alone. 

At the meeting of the American Association for the Advance- 
ment of Science, at Brooklyn, in August last, I proposed two new 
methods of mapping the corona without an eclipse. It is the 
object of the present paper to describe the more important one 
of these methods. 

When a bolometer is used with a reflecting galvanometer, it 
is well known that the galvanometer deflection is always pro- 
portional to the intensity of the radiation to which the bolometer 
is exposed. A moment's consideration will show the importance 
of this fact in the present connection. Let us suppose that an 
image of the Sun and its surroundings has been formed by a 
reflecting telescope, and that we wish to determine the radiation 
from the region lying outside the Sun's limb with a bolometer. 
Call 5o the deflection obtained when the bolometer is exposed to 
the sky at a distance D from the limb {D being greater than the 
radius of the corona) ; Cj^ , the deflection that would be given 
by the corona alone with the bolometer at the distance R from 
the limb ; S^^ , the deflection that would be given by the sky 
alone at this point. Then if a differential bolometer is used, 
with one member at the distance A and the other at the distance 
7?, from the limb, the deflection will be 

Thus the absolute intensity of the sky radiation is of little 
importance ; the value of {Sj^—Sj,) measures the disturbing effect. 
During a total eclipse 5]^ differs Httle from 5^ , and the radiation 



320 GEORGE E. HALE 

measured would be in large part due to the corona. It is evident 
that D should be made as small as possible. When good con^ 
ditions prevail at great altitudes above the earthy the brightness 
of the sky does not increase very rapidly as the Sun's limb is 
approached. In this case {Sj^—S^) would be small, and the con- 
ditions of a total eclipse would be approximated. In any case 
the conditions are the same as though the sky were black up to 
a distance D from the Sun's limb. It can therefore hardly be 
doubted that the bolometer or radio-micrometer could be 
employed so as to indicate the coronal radiation, and to distin- 
guish it from that of the Earth's atmosphere. 

That this opinion is well-founded will appear from a closer 
examination of the question. Unfortunately our knowledge of 
the heat radiation of the corona is very slight, but two independ- 
ent ways of treating the matter lead to results of the same order 
of magnitude, which will suffice for the present purpose. 

In the first place let us consider the intensity of the corona's 
light, as determined by photometric measures made during total 
solar eclipses. Prior to 1878 the brightness of the corona was in 
most cases estimated by the observer, and it is not surprising that 
the conclusions differ widely among themselves. An examina- 
tion of the observations made during total solar eclipses so care- 
fully compiled and discussed by the late Mr. Ranyard' shows 
that few of the earlier estimates can be considered trustworthy. 
In i860 Mr. J. M. Wilson was able to see the prominences and 
the lower part of the corona through the base of a wedge of dark 
glass, which extinguished the light of the full Moon at about its 
middle part.* At the eclipse of July 29, 1878, Professor J. W. 
Langley made some measures of the corona's light at his station 
on Pike's Peak. At a distance of I ' from the Moon's limb the 
brightness was found to be six times that of the surface of the 
full Moon, while at a distance of 3' the brightness was only one- 
sixtieth of that previously measured.' These observations, and 

*Mem, R,A,S. 4X. 

*Loc, cii. p. 251. 

9 Wash, Obs. 1876, Appendix 3, p. 214. 



METHOD OF MAPPING THE SOLAR CORONA 121 

others on the total light of the corona made during the same 
eclipse, were discussed by Professor William Harkness. His 
conclusions were as follows : 

*' I. The total light of the corona was 0.072 of that of a stand- 
ard candle at one foot distance, or 3.8 times that of the full 
Moon, or 0.0000069 of that of the Sun. 

''2. The photographs show that the coronal light varied 
inversely as the square of the distance from the Sun's limb. 

" 3. The brightness of any part of the corona is given, quite 
approximately, in terms of the brightness of the surface of the 
full Moon, by the expression — 

^ = ^(23'+ 100' cos ^) 

(where h is the distance from the Sun's limb in minutes of arc, 
and ^ the latitude, measured from the Sun's equator, of the part 
in question) . For very small values of h this formula fails. Prob- 
ably the brightest part of the corona was about fifteen times 
brighter than the surface of the full Moon, or 37000 times fainter 
than the surface of the Sun. 

"4. The corona of December 22, 1870, seems to have been 
seven and one-fourth times brighter than that of July 29, 
1878." 

The last conclusion was reached by a comparison of the 1878 
measures with the results of some photometric observations made 
at Jerez, Spain, in 1870, by Mr. W. O. Ross, under the direction 
of Professor E. C. Pickering. 

Professor Harkness is careful to point out ''that, in deducing 
these results, visual and photographic data have been intermin- 
gled as if they were homogeneous. Such a procedure is cer- 
tainly open to objection, but it is not likely that it introduced 
any error grave enough to impair the value of the results as first 
approximations.' 

Subsequent investigations on the brightness of the corona 
have not borne out Professor Harkness' conclusions. Captain 
Abney's discussion of the results obtained at the eclipse of 1886 

* Wash, Ois, 1876, Appendix 3, p. 392. 



322 



GEORGE E. HALE 



is adverse to the law of inverse squares ;* and Professor W. H. Pick- 
ering's photographic photometry at the same eclipse showed the 
falling off in intensity to be much less rapid than Professor 
Harkness' formula requires.' Nor do the measures made on the 
negatives obtained by the Lick Observatory parties at the 
eclipses of January and December, 1889, decrease with such 
rapidity in going outward from the Sun's limb. The following 
tabled brings together some of these results for comparison: 





Pickering, 


Holden. 


Holden, 




Augostt 1886 


Janiiaiy, 1889 


December, 1889 


Intrinsic actinic brilliancy of the brightest 








parts of the corona 


O.03X 


0.079 


0.029 


Intrinsic actinic brilliancy of the polar rays 








Tabout) 




0.053 


0.016 


Intrinsic actinic brilliancy of the sky near 








corona 


0.0007 


0.0050 




Ratio of intrinsic brilliancy of the brightest 








parU of the corona to that of the sky 








(actinic) 


44 to I 


16 to X 


32 to I 


Intrinsic actinic brilliancy of the sky at i "* 








from the Sun in daylight (average) 


40 


• • • . 


• • • • 


Ratio of intrinsic brilliancy of the brightest 








parte of the corona to thai of sky at i° 








from the Sun in daylight (about) 


I to X290 


I to 506 


I to 1379 


Intrinsic actinic brilliancy of the full Moon 


X.66 


• . . . 




Ratio of intrinsic brilliancy of the brightest 








parts of the corona to that of full Moon. 


X to 53.6 


I to 2X 


X to 57.2 



The differences between the values obtained for the bright- 
ness of the corona at these eclipses are very striking. Professor 
Harkness considered the brightness of the brightest parts of the 
corona to be about fifteen times that of the full Moon, while 
the results obtained in 1 886, and in January and December, 1 889, 
w^^^ T^T* tIt ^"^ T^T respectively, if the brightness obtained 

^Phil, Trans. (A.) 1889, p. 380. 
^ Harvard Annals^ x8, 100. 

3 £. S. Holden : Lick Observatory Report on the Observations of the Total Edipse 
of the Sun, December ai-aj^ i88g, p. 14. 



METHOD OF MAPPING THE SOLAR CORONA 323 

by Professor Harkness be taken as unity. Moreover, Professor 
Harkness concluded that the corona of 1870 was 7^ times 
brighter than that of 1878. * 

A number of causes might conspire to produce these remark- 
able differences. The duration of totality, altitude of the Sun 
and consequent variation in atmospheric absorption, clouds or 
haze during totality, and differences in the methods employed 
are all to be considered, in addition to changes in the corona 
itself. 

1878 was a year of minimum Sun-spots, and at the time of 
the eclipse spots and prominences were few and small. The 
photometric measures of Professor Langley were made on Pike's 
Peak (14,147 feet) on July 29, when the Sun's altitude was over 
40**. The duration of totality was about 2^"^. During the entire 
eclipse the sky was perfectly clear, and of a deep blue color. 
The eclipse of August 29, 1886, occurred during a Sun-spot 
maximum. Professor Pickering's photographs were made at 
sea-level in the exceedingly moist atmosphere of the island of 
Grenada. The duration of totality was about 6^°*, so that the 
lowest and brightest parts of the corona were covered most of 
the time. The observations were made through passing clouds, 
and after 1 56 seconds of totality the corona was covered with 
hazy clouds, which rapidly grew denser, so that nothing could be 
seen 24 seconds later. The Sun was only 20° above the horizon 
during the eclipse. Most of the photographic plates were ruined 
by the moist atmosphere, and the photometric measures made on 
those that were preserved are surely subject to some uncertainty. 

The two eclipses of 1889 came in a period of minimum solar 
activity. On January i there were clouds before and after the 
eclipse, and during totality the corona was seen through a slight 
haze. The altitude of the Sun was 24° 6', and the duration of 
totality about 2™. The station of the Lick Observatory party 
was 2040 feet above the level of the sea. 

In December the conditions under which the Lick party 
observed were somewhat similar to those experienced by Pro- 
fessor W. H. Pickering in 1886. The station at Cayenne was 



324 GEORGE E. HALE 

little above the level of the sea, and the climate was very damp. 
The altitude of the Sun was about the same as in the eclipse of 
January. Just before second contact a heavy rain fell, and dur- 
ing totality the conditions were not at all favorable. The plates 
were developed soon after the eclipse, and it was found that the 
extreme moisture had not affected them appreciably. They 
were all greatly over-exposed. 

The Carcel burner used to standardize the plates from which 
the photometric results in the table for the eclipses of 1 889 were 
deduced, was found to be constant for an hour or more, but 
"enormous variations occur between the results of different days 
(so that no results can be drawn from a comparison of plates 
standardized at different times)."* Professor Holden expressly 
states that until the reductions can be repeated the results are to 
be regarded as provisional. 

It would appear from a comparison of these results that the 
determinations of the brightness of the corona made by Profes- 
sors Pickering and Holden are in some respects less satisfactory 
than the values given by Professor Harkness, though it is not 
probable that the brightness falls off so rapidly as the law of 
inverse squares requires. Professor J. W. Langley's observations 
seem to me quite sufficient to prove that the insual brightness of 
the corona and sky at a distance of i ' .66 from the Sun's limb 
was at least as great as that of the surface of the full Moon. 
This differs so widely from the photographic results that one is 
led to suspect the existence of a maximum of intensity in the 
lower part of the coronal spectrum. In spite of the meagerness 
of the data at our disposal it is possible to derive conclusions 
of some value in this connection. 

The few observations made prior to the eclipse of 1878 
seemed to show that the polarization of the light of the corona 
increased toward a maximum in passing outward from the Moon's 
limb.* In an important theoretical discussion, published in 

* A. O. Leuschner : Lick Observatory Report on the Total Eclipse of December 
Mi-22, t88g^ p. 9. 

•Ran yard: Mem. R. A, S. 4X1 260. 



METHOD OF MAPPING THE SOLAR CORONA 32$ 

1879,' Professor Arthur Schuster assumed various laws of dis- 
tribution of the polarizing particles in the corona, and applied 
his conclusions to the observations of Winter at the eclipse of 
1 87 1. Mr. Winter found the polarization to be greater at a point 
about 10' from the Moon's limb than at another point closer to 
the limb.* The proportion of polarized light to the whole light in 
the former position he determined to be about 0.246. Professor 
Schuster, while not attaching great importance to these results, 
assumed the maximum of polarization to be at just the point 
observed by Mr. Winter, and calculated the amount of light due 
to scattering matter uniformly distributed in the corona to be 
over 90 per cent, of the whole. This assumption of uniform 
distribution he pointed out to be very improbable, as, if it were 
true, the light not due to scattering particles would increase at 
an extremely rapid rate in going out from the limb. If, on the 
other hand, the density of the scattering matter varied inversely 
as the square of the distance from the Sun's center, only about 
one-half of the light would be due to it. With this law of dis- 
tribution the light not due to scattering would decrease outward 
from the limb. Professor Schuster concluded that at the eclipse 
of 1 87 1 the intensity of the scattered light was probably little 
more than one-half of that of the total light.3 

At the eclipse of July 29, 1878, Professor Arthur W. Wright 
made a very thorough visual and photographic investigation of 
the polarization of the coronal light. None of the photographs 
gave any trustworthy evidence of a region of maximum polariza- 
tion at a distance from the limb. On the contrary, many of the 
plates showed a distinct increase in the intensity of polarization 
near the limb. With the polarimeter the value of the polariza- 
tion between 4' and 10' from the Moon's limb was found to be 
nearly 12 per cent.; between 12' and 18' it was 4.7 per cent; 
at 22' there were traces of polarization, but the light was too 
feeble to permit of measurement. The polarization was nearly 

^M. N. 40, 35. 

•Mem. R, A. S. 41, 324. 

iM. N. 40, 54. 



326 GEORGE E. HALE 

uniform around the circumference, except in regions extending 
20^ on each side of the poles, where it was somewhat greater. 
The observations did not seem to be affected by the passage of 
polarized rays through the Earth's atmosphere. 

In his concluding remarks Professor Wright offers an expla- 
nation of the results. From the fact that the polarization is 
radial, it is clearly due to reflection of solar light from matter in 
the corona. Finely divided matter, such as that of vapors just 
in the act of condensing, polarizes light completely at an inci- 
dence of 45*^. On account of the intense heat of the Sun, mat- 
ter in this form would very likely be found near the surface, 
while farther out in the corona larger particles would be more 
numerous. For these particles the polarization would increase 
outward from the Sun. But as they approached the photosphere 
they would be greatly reduced in size, or even vaporized ; and in 
this condition their polarization would probably be sufficient to 
produce the effect observed. "The inner portions of the corona 
must be of sufficiently high temperature to be self-luminous, a 
circumstance which would diminish the apparent percentage of 
polarization, and in part explain the smallness of its amount. As 
the light thus emitted must come chiefly from matter in the solid 
form, its spectrum would be continuous, in the absence of any 
atmosphere sufficient to produce dark absorption lines. More- 
over, as this light is superadded to that which is reflected, it 
would diminish the intensity of the dark lines of the spectrum, 
caused by the latter, and render them more difficult of observa- 
tion — a fact in harmony with experience."* 

The conclusion thus deduced from polariscopic observations 
of the existence of incandescent particles in the coronal atmos- 
phere is strengthened by a consideration of the temperature 
conditions which must exist in the corona. Mr. Ranyard 
pointed out in 1891' that experiments with great "burning 
glasses '' are sufficient to show that the temperature in the coronal 
region must be high. Platinum would certainly melt at distances 

* Arthur W. Wright: Wash, Ods, 1876, Appendix 3, p. 280. 
^Knowl, 14, 14. 



METHOD OF MAPPING THE SOLAR CORONA 3^7 

but little less than a solar radius. Many of the glowing par- 
ticles in the lower regions of the corona are thus in all probability 
liquid ; but the most refractory substances may also be present 
in the solid state. 

It is evident that the scattering effect of the smallest particles 
in the corona must be most marked for light of short wave- 
length. If most of the particles within a certain radius are small 
in comparison with a wave of blue or violet light, the scattered 
light, could it be observed alone, would probably exhibit a max- 
imum at the upper end of the spectrum. But we have just seen 
that the scattered light is only a part of the total radiation of 
the corona. The maximum of intensity in the continuous spec- 
trum of the glowing particles must be displaced toward the red 
as compared with the spectrum of the photosphere, as the par- 
ticles are at much lower temperature. The absorption in the 
upper part of the solar spectrum tends to produce a similar shift 
of the maximum ; and the spectrum of the scattered light must 
be added to that of the luminous particles to give the true cor- 
onal spectrum. Hence, in spite of the lower temperature of the 
corona and the fact that the coronal radiation is not subject to 
the general absorption, the maximum may not be much lower 
than that of the spectrum of the photosphere.' 

We may safely conclude that the coronal spectrum is of con- 
siderable intensity in the yellow, red, and infra-red, and that its 
heat radiation should be easily measurable. It is strange that 
but few attempts have been made to measure this radiation dur- 
ing total eclipses. In 1842 Magrini found that the heat of an 
image of the corona formed by a reflecting telescope was such 
as to move the index of a Rumford thermoscope half through 
the scale by the end of totality, while the heat from the full 
Moon could not be detected with the same apparatus. Unfor- 
tunately, Magrini's description of his work does not give all of 

' For various reasons the observations that have been made on the relative intensity 
of different parts of the coronal spectrum can hardly be depended upon to decide this 
question. The photographs are quite as unreliable, as platto equally sensitive to all 
parts of the spectrum have never been employed. 



328 GEORGE E. HALE 

the details, and it is impossible to assign a proper weight to the 
result.' In i860 Sir William Thomson (now Lord Kelvin) pro- 
posed that the heat of the corona be measured by spirits of 
wine or mercury thermometers of the ordinary form, and pointed 
out that a mirror would be better than a lens for this work 
" because the hot dust around the Sun must produce radiant heat 
of such colour as that of a hot stone or metal not at a bright red 
heat."' This suggestion does not seem to have been acted upon, 
but in 1878 Mr. T. A. Edison exposed his recently invented 
"tasimeter" to the total light of the corona, and the galvanom- 
eter needle was thrown off the scale.' During the same eclipse 
Professor Young placed a thermopile in the infra-red region of 
the coronal spectrum, and obtained a doubtful indication of a 
heat band.^ 

So far as I am aware the observations of Magrini and Edison 
are the only ones hitherto made on the heat radiation of the 
corona. Taken in connection with the results obtained for the 
brightness, they certainly seem to show that the bolometer ought 
to indicate a greater intrinsic radiation from the corona than 
from the full Moon. 

In the following discussion of the sensitiveness of heat- 
measuring instruments the bolometer alone will be considered, 
as for it the required data are most readily obtainable. The 
experiments of Boys and Paschen seem to show that the radio- 
micrometer may also be advantageously employed, and the 
thermopile or selenium cell might be modifieti to meet the special 
requirements of the investigation. In all of my preliminary 
experiments I have used bolometers exclusively. 

In his Prize Essay On the Distribution of the Moon's Heat and 
its Variation with the Phased Professor Frank W. Very has given 
data which are valuable for our purpose. The bolometer 

* Ranyard : Mem, H. A, S, 41, 246. 

* " On the importance of making Observations on Thermal Radiation during the 
coming Eclipse of the Sun." M, N, ao» 3X7* 

^Am.J,ScL X17, 52. 

* Princeton Review, 1878, 884. 

5 Published by the Utrecht Society of Arts and Sciences, The Hague, 1891. 



METHOD OF MAPPING THE SOLAR CORONA 3^9 

employed exposed a sensitive surface of about 19 square milli- 
meters. The image of the Moon, about 28"™ in diameter, was 
formed by a silver-on-glass mirror, the aperture of which is not 
mentioned. With the galvanometer used, a small region near 
the center of the disk of the full Moon gave a deflection of 
nearly loo"^ scale divisions. 

With similar apparatus we might expect to obtain at least 
this deflection for the brightest parts of the corona, if the bolom- 
eter were used differentially, and S^ were equal to S^. As Sj^^S^, 
the deflection should be somewhat greater. 

It is evident, however, that if we wish to map the corona, so 
as to show its structure, a much higher "resolving power" will 
be required. To secure this, the bolometer must be made much 
smaller, and its greatest dimension should be radial to the Sun. 
With a 28"^ solar image the bolometer strip might perhaps be 
2mm long (radially) and 0"".5 wide. As much of the coronal 
structure is radial, or nearly so, such a bolometer would be small 
enough for a preliminary investigation. Its area would be only 
^ of that used by Professor Very, and with an equally sensitive 
galvanometer the deflection due to the corona alone might not 
exceed 5 or 6 divisions. Fortunately, however, a galvanometer 
about 75 times as sensitive as the Allegheny instrument may 
now be constructed, and its deflection with the small bolometer 
should be not less than 350 or 400 divisions. With good con- 
ditions of steadiness and freedom from magnetic disturbances, a 
deflection of \ division can be determined with certainty. A 
variation in the coronal radiation of one part in 1 500 should there- 
fore be measurable.* If such a bolometer were exposed to 
various parts of the image, the general structure of the corona 

' The differential method to be employed will reduce the effect of variations in the 
radiation of the EarthV atmosphere to a minimum, as in most cases both bolometers 
will be affected alike. Irregularities in the curve not due to the corona can be elimi- 
nated by repeated experiments. The effect of increase in brightness of the sky 
toward the Sun's limb will he shown in a gradual change in the length of the mean 
ordinate of the curve. The resulting effect upon the appearance of the coronal image 
will be analogous to that produced by the brightness of the sky and diffraction at the 
Moon's limb during an eclipse. (See Barnard : Lick Observatory Report on the Total 
Eclipse of January /, tB8<), pp. 65, 69; Keeler: Ihid.^ p. 45). 



330 GEORGE E. HALE 

could be made out in a manner similar to that employed in 1868 
to determine the form of prominences by means of visual 
observations of the Ha line with a narrow slit. 

Although results of some value could probably be obtained by 
this process, yet it is evident that a much better method is desirable 
for the purpose of making daily records of the form and structure 
of the corona. A means of accomplishing this would be to move 
the bolometer over the coronal image (or the image over the 
bolometer) in a manner similar to that employed by Professor 
Langley in his spectro-bolographic work. A photographic plate 
upon which the spot of light from the galvanometer mirror falls 
is moved synchronously with the bolometer (or image) at right 
angles to the direction of the deflection. A curve is thus photo- 
graphed, whose ordinates measure the heat radiation to which 
the bolometer was exposed in the corresponding positions on the 
image. It is perhaps simplest to consider the bolometer to be 
moving in a circle just outside the Sun's limb. After it has com- 
pleted one revolution, it is moved out radially a distance equal 
to its own length, and is then moved through another circle con- 
centric with the Sun. The photographic registration is mean- 
while continued with no interruption other than the momentary 
one due to the shifting of the bolometer. This process is repeated 
until the entire corona has been traversed by the moving bolom- 
eter; the other bolometer of the pair being maintained in a 
fixed position on the image of the sky at a distance of from 
30' to 100' from the Sun's limb.' The result will be a nearly 
continuous photographed curve showing the heat radiation from 
all points in the corona. 

Instead of being moved in a straight line the photographic 
plate may be rotated about its center. In fact, the bolometer 
may be on one side of a disk, and the sensitive plate on the 
other. This extremely simple arrangement would render com- 
plicated synchronizing apparatus unnecessary. A slight modifica- 
tion of the same plan would allow rectilinear motion of the plate. 
But if the plate were made to rotate, we would have a series of 

' Preferably in the region of the pole. 



METHOD OF MAPPING THE SOLAR CORONA 331 

closed curves of increasing radii, whose maxima and minima 
would enable us to determine at a glance the general form of the 
corona. The distortion of the curves could easily be removed 
if necessary. 

It will be advantageous, however, to go a step further in 
order to obtain an image of the corona comparable with an 
eclipse photograph. Suppose a photographic plate to be placed 
behind a screen which is illuminated by parallel light from a 
constant source. In the screen is an opening equal in length to 
the bolometer strip and variable in width. The screen is mounted 
upon an axis opposite the center of the sensitive plate, and the 
radius of the circle described by the opening when the screen is 
rotated is made equal to the radius of the circle described by the 
bolometer in its motion over the coronal image at the focus of the 
telescope. The width of the slit in the screen is controlled by 
the galvanometer curve, which is cut out of pasteboard or other 
suitable material. It is evident that if the screen is rotated at 
a uniform speed before the plate and the curve moved at a cor- 
responding rate over the controlling mechanism, the intensity of 
the photographic action at any point will be nearly pro- 
portional (or, by a slight change, inversely proportional, if a 
positive is required) to the ordinate of the curve at the corre- 
sponding point. After the first circle has been photographed, 
the sliding strip which carries the opening in the screen is moved 
out through a distance corresponding with the length of the 
bolometer strip, and the process is repeated with the second sec-« 
tion of the curve. Continued repetition of this procedure will 
give a negative (or positive) of the corona. 

Professor F. L. O. Wadsworth, to whom I am indebted for 
many valuable suggestions, has pointed out that the screen 
might be fixed at a distance, with a lens interposed to form an 
image of the uniformly illuminated opening (equal in size to the 
bolometer strip) upon the photographic plate. The lens is 
covered by a diaphragm which exposes only a narrow strip 
across the center. The opening in the screen is placed at a 
distance from the optical axis of the lens equal to the radius of 



332 GEORGE E. BALE 

the circle described by the bolometer in its motion around the 
Sun. The photographic plate is made to revolve at a uniform 
speed about the optical axis of the lens, and the galvanometer 
curve, cut out of stiff material, is moved synchronously across the 
diaphragm in the lens, at right angles to it. The brightness of 
the spot of light upon the moving plate should thus be 
inversely proportional to the ordinate of the curve at the point 
where it crosses the diaphragm. Repetition of the process will 
gradually build up a positive image of the corona. 

Many other methods of accomplishing this result will suggest 
themselves. A very simple one has recently occurred to me. 
In the first method described above we may replace the photo* 
graphic plate by a sheet of white paper, and employ, instead of 
parallel light, one of the ''air brushes" used by photographers in 
working up bromide enlargements. In this ingenious little 
device a stream of air, carrying with it finely divided India ink, 
is thrown upon the paper. The flow is regulated by pressure of 
the finger upon a delicate valve. For our purpose the air brush 
is supported over the opening in the screen (which is made 
equal in size to the bolometer strip), and the flow of ink is 
regulated by the galvanometer curve, which is cut out of stifle 
material, and moved over the valve.* 

It may be urged that the methods suggested are too slow to 
be of practical value, especially as the bolometer should move 
over its own width in the time of swing of the galvanometer 
• (for this work about ten seconds). To obviate this, several 
bolometer strips may be mounted end to end, so as to form a 
compound bolometer extending out from the Sun's limb to a 
distance of from 20' to 30'. All the bolometers would be 
differential, as before, the second member being placed outside 
the corona. As many galvanometers as there are bolometers 
would be required. In transforming the curves into an image 

'The bolometer might also be moved over the corona in a spiral path or in 
straight lines ; the transformation methods may be easily modified to conform to these 
conditions. 

A little consideration will show that the galvanometer might be made to record an 
image of the corona directly upon a photographic plate, if this were thought desirable. 



METHOD OF MAPPING THE SOLAR CORONA 333 

of the corona any of the processes suggested may be used, with 
one or several openings in the exposing screen. 

The time required for the operation might be still further 
decreased by using another form of compound bolometer, con- 
sisting of a large number of short radial strips mounted in two 
parallel planes, separated by a thin air space. The members 
could thus be made to overlap slightly, so that the entire corona 
would be covered with bolometers. All of these bolometers 
would be used differentially with one or several galvanometers. 
The battery current would flow through them constantly, and 
the galvanometer circuits would be completed by a system of 
rotating contacts. As each bolometer would be constantly 
exposed to the coronal radiation, the difficulties due to exposure 
of a single strip to rapidly varying heat conditions would in part 
be obviated, with a consequent gain in the rapidity of the process. 
The chief objection to the compound bolometer in either form is 
the impossibility of making all the strips of exactly the same size 
and resistance. Professor Wadsworth has, however, suggested 
methods by which this difficulty may be in large part overcome. 

In my experiments at the Kenwood Observatory I have 
carried the preliminary investigations as far as the conditions 
will allow ; they have dealt chiefly with the construction and 
testing of various forms of phosphor-bronze, platinum and steel 
bolometers.' The location of the Observatory is quite unsuitable 
for work with the delicate apparatus required. The galvanometer, 
which is being constructed under Professor Wadsworth's super- 
vision, promises to be even more sensitive than the extremely 
delicate instruments designed by him for the spectro-bolographic 
work of the Smithsonian Astrophysical Observatory.' Such a 

' The bolometers were mounted in a brass tube, provided with a large number of 
diaphragms, each of which was blackened on the lower and polished on the upper 
surface. This tube was soldered within another brass tube of much larger diameter, 
with double walls at the end. Water was constantly passed through the chambers at 
the end and sides ; the bolometers were thus screened from the radiation of the room, 
and completely protected from the intense heat of the direct solar image. Only the 
heat of the corona and sky could reach them through the small openings of the 
diaphragms. 

■ See the February number of this Journal, p. 163. 



334 GEORGE E. HALE 

galvanometer cannot be used in the midst of a large city. The 
smoke is another obstacle, as it increases the atmospheric absorp- 
tion, and also the value of (5^, — S^). The special apparatus for 
the investigation will probably be completed by the end of May, 
when it is expected that the instruments of the Kenwood 
Observatory will be transferred to the new Yerkcs Observatory 
at Lake Geneva, Wisconsin. There the conditions for continu- 
ing the work on the corona could hardly be surpassed, especially 
at the time of the summer solstice. 

It is evident that the method described in this paper may be 
applied to many purposes other than that for which it was pri- 
marily designed. The bolometer, radio-micrometer or thermopile 
will be used, as circumstances may require, in a number of 
investigations to be undertaken soon by myself and my assist- 
ants. The heat-measuring apparatus will be used alone, or in 
connection with a spectroheliograph, in attempts to obtain ''heat 
images" of the corona, prominences, spots, faculx and other 
solar phenomena, and also of the Moon. Other investigations 
depending upon this method, or closely allied to it, will be 
described in a future paper. 

Kenwood Observatory, Chicago, 
March 15, 1895. 



ON A NEW FORM OF SPECTROSCOPE.' 

By C. PULFRICH. 

In the construction of the spectroscope which is described 
below, an attempt has been made, and I believe for the first time 
in this form, to apply the method of normally reflected rays, or 
principle of reversed path, to the well-known compound prisms 
of Rutherfurd (a flint-glass prism of large angle cemented between 
thin prisms of crown glass) and of Wernicke (ethyl cinnamate 
instead of the flint glass). In doing this I believe that I have 
obviated certain difiiculties attending the use of these prisms 
(which are so favorably known on account of their high disper- 
sion and transparency) in previous constructions, and have made 
them more suitable for spectroscopic and perhaps also for spec- 
trographic' studies than they were before. 

The advantages of the method of normally reflected rays, 
as applied to the spectroscope, lie chiefly in the two following 
points — to which may be added the less important one of con- 
venience in observation, inasmuch as the different parts of the 
spectrum are observed without changing the position of the tel- 
escope or photographic plate : first, the position of minimum 
deviation is given to the system of prisms by a comparatively 
simple arrangement, so that any spectral line in the center of the 
field of view is necessarily seen under the conditions of maximum 
distinctness, and second, the focal length of the collimator and 
observing telescope can be made relatively very great without 
making the apparatus inconveniently large for practical use ; other 
circumstances being equal, a considerable increase in purity and 
extension of the spectrum is in this way obtained.^ 

Let us first consider somewhat more closely the forms of 

^'Yr%mi\9X^bKim^<t ZieUschrifi fiir Inshmwuntenkunde, 10 Heft, 1894. Com- 
mnnication from the optical establishment of Carl Zeiss, in Jena. 
'Compare Lohse, Z/. Imtrum, 1885, p. 11. 

'Compare Lippich, Centrai-Zeiiung fur OpHk und Mech., 188 1, pp. 49 and 61. 

335 



336 C. PULFRICH 

spectroscope to which the principle of reversed path has already 
been applied. 

The first application of the principle to the construction of a 
spectroscope with only a single prism seems to be due to Duboscq, 
who followed the designs of v. Littrow.' In front of the tele- 
scope, which was fixed, and provided with some form of slit, 
(how the slit was arranged I have not been able to ascertain) 
was a 30° prism, capable of rotation about an axis M (Fig. i).* 

The back surface of the prism was silvered. Duboscq recog- 
nized the features which are characteristic of his method, namely : 
that the length of path of a ray returning upon itself is exactly 
the same as if it passed through a prism of twice the angle 




Fig. I 

(or 60°) in the position ot minimum deviation; further, that a 
simple rotation of the prism sufficed to bring the di£ferent colors 
into the field ; and finally that, other circumstances being equal, 
the dispersion is exactly the same as in the case of an entire 
prism. It is, therefore, somewhat surprising that this Duboscq 
spectroscope should be almost unused in laboratories, while the 
ordinary form of spectroscope with 60° prism is so common. 

The first application of the same principle to several prisms 
of 60° each, placed so as to form a train, is due to v. Littrow. 
The (four) prisms could be so moved, by means of a special 
mechanical device, that each prism remained in the position of 
minimum deviation for the normally reflected and observed ray. 
The path of a ray is shown in Fig. 2 sufficiently well for the pur- 

"V. Littrow, Wien, Ber, 47, II, 29, 1863. 

* In principle, therefore, exactly the same arrangement as in the spectrometer by 
Professor Abbe, which is so admirably adapted to measurements of the refractive and dis- 
persive powers of a body. (Z/. Instrum. 1889, p. 361, and KiUalog ubtropiischt Mess- 
instrumente von Carl Zeiss^ Jena, 1893.) 



A NEW FORM OF SPECTROSCOPE 



337 



pose of illustration. The resulting dispersion is in this case also 
equal to that of twice the number of prisms when the ray is trans- 
mitted. Littrow also mentions the plan, later repeatedly carried 
into execution,' of increasing the dispersion fourfold by means 
of a special reflecting prism, which transferred the rays to another 
story before they were finally sent back on the same path. 

Following V. Littrow, the principle of reversed path was used 
in the construction of spectroscopes by many persons (Browning, 




.. i\ 



r-V 



Fig. 2 

Hilg<?r, Kruss, Grubb, Brackett and others).' The device for 
keeping the prisms in the position of minimum deviation has 
been gradually perfected in these different forms of apparatus.^ 
In another respect also these now very popular forms of spectro- 
scope have been brought to a continually increasing state of per- 
fection ; the arrangement by which the reflecting half-prism (or 
mirror) can be introduced between any two prisms of the train is 

' Compare, among others, Cornu, Z/. Insirum, 1883, P* I7I- 

'Compare Kriiss, " Uebcr Spektralapparate mit automatischer Einstellung." Z. /. 

Instrum. 1885, pp. 181 and 232; x888, p. 388. 

3 See in particular the ingenious arrangement recently described by Kriiss. {Z. /. 

Instrum. 1890, p. 97*) 



338 C. PULFRICH 

especially worthy of mention. It enables the observer to alter at 
will the dispersion of his apparatus, within certain limits deter- 
mined by the number of prisms in the train, without in any other 
way changing the disposition of the parts of the instrument.' 

With regard to the application of the principle of reversed 
path to the compound prisms of Rutherfurd or Wernicke, we 
have here two different ways, one of which has already been 
taken by Kriiss,' Prazmowski,^ and Schmidt and Haensch,^ and 
which consists in either wholly or partly replacing the 60° prisms 
of the automatic apparatus already described by compound 
*prisms. As compared with 60^ prism spectroscopes of equal 
dispersion these spectroscopes have at least the advantage of 
giving brighter spectra, on account of the smaller number of 
internal reflections. 

The second, and as it seems to me the more direct way, is that 
which I have taken in constructing the spectroscope described in 
this paper. It will be seen from the following description that 
the construction is closely analogous to that of Duboscq. 

We may regard every Rutherfurd or Wernicke prism, whether 
it consists of three or of more than three component prisms, as 
divided by a plane passing through the refracting edge of the 
central prism, into two parts, which are symmetrical with respect 

' A new form of Littrow spectroscope, differing somewhat from the above, is 
described by Mr. F. L. O. Wadsworth in the July (1894) number of iht PkiUsofihual Mag- 
atine. The special feature of this apparatus is that the objective is replaced by a sil- 
vered concave mirror of about 1.7 meters (or another of 4.7 meters) focal length, with 
the object of entirely avoiding the annoying reflecttons produced by the surfaces of an 
objective. The rays coming from the slit are thrown by a small totally reflecting 
prism upon the concave mirror, which reflects them as a parallel bundle to a 60* prism 
of flint glass, behind which is a plane mirror. The same concave mirror unites the 
returning rays to form an image of the spectrum. This ingenious apparatus would 
probably be considerably improved if the prism and plane mirror were replaced by the 
combination which is described further below in the present article. I have been 
informed by a personal communication that Messrs. Kayser and Runge have con- 
structed an apparatus quite similar to that of Wadsworth, with which they intend to 
photograph spectra in vacuo, 

« Krilss, Z/. Imtrum. 1887, p. 183. 

3 Prazmowski, Z / Imtrum, 1889, p. 106. 

4 Schmidt and Haensch, Speziai-ICataiogt Aussteilung in Chicago^ 1893, P« 70* 



A NEW FORM OF SPECTKOSCOPE 339 

to the plane (Figs. 3 and 4). Each half, when the rays are 
returned on their path by reflection, is equal in dispersive power 
to the original prism with transmitted rays. It is also evident 
that the automatic adjustment for minimum deviation is obtained 
by rotating the prism about the axis M, exactly as in the case of 
the simple prism illustrated in Fig. i. 

This arrangement also enables us to make the spectroscope 
equal in dispersive power to any number of 60 ° prisms, without 
obliging us to use any other device for automatic adjustment 
than the simple axis of rotation. This axis not only answers the 
required purpose in the most perfect manner, but it has a still 
further advantage; the apparatus for measuring the distances 
between spectral lines (graduated circle and micrometer micro- 




scopes) can be attached directly to the most rigid part of the 
instrument, — the axis of rotation, — in consequence of which a 
much higher degree of accuracy can be reached in the measure- 
ments than with other methods of construction. 

It will be readily understood that practical reasons prevent us 
from indefinitely increasing the dispersive power of a compound 
prism by increasing the number of component prisms. The 
technical difficulties of construction in doing this grow at a 
greatly increased rate. A certain limit is thus set to the increase 
of dispersion, which can be exceeded only at the cost of the 
purity of the spectrum. Viewed in this light, the construction 
which we are considering is very favorable, inasmuch as we 
have to do with only half of the prism. We require therefore 
only half of the material, half the number of cemented surfaces, 
and what is also of importance in the technical production, we 
do not require the prism angles on opposite sides of the plane of 



340 C PULFRICH 

symmetry to be rigidly equal, as they must be in the ordinary 
construction. This condition is necessarily fulfilled with abso- 
lute exactness. 

It is at least certain that the production of compound prisms 
consisting of only two or three single prisms does not present 
any difficulty worth considering. We shall illustrate the action 
of such prisms, made up of two or three parts, by means of a few 
examples. 

The prism represented in Fig. 3 consists of a prism of flint 
glass combined with one of crown glass (both Jena glass of 
relatively very great transparency), in which the highest degree 
of dispersion that can be obtained with a double prism is pur- 
posely avoided, and in fact is not nearly approached.' The 
combination is remarkable for the brightness of the spectrum in 
the violet. The optical constants of the glass (refraction and 
dispersion ) expressed in the usual manner, are as follows : 
Flint glass prism AEF^ refracting angle = 57® 

«/,^= 1.6800, Hg: — «£: = o.o2i52, r = 3i.6; 
Crown glass prism AFD^ refracting angle = 40® 

«/>=i«5i70» ^F — «c=o-oo847, r = 6i.o. 

From these data the dispersion, 1. e., twice the difference of 
direction of the rays normally reflected on the back surface 
(twice the angle of rotation of the prism) is by computation 
4** 34' between C and F. Taking the dispersion of a single 60® 
prism of ordinary flint glass («/, = i.62) for the same lines, as 
I J^**, the combination represented in Fig. 3 is equal in disper- 
sion to three such prisms. 

A correspondingly greater dispersion can be obtained by 
using three prisms cemented together. It will in that case be 
equal to the dispersion of a Rutherfurd prism with five members, 
and can be made equal to the dispersion of six, seven or eight 
single prisms of 60**, according to the optical constants of the 
glass used in the construction of the prism. 

We see therefore that it is not difficult, by using only two or 

' The spectroscope furnished with this prism ( see further below ) shows the nickel 
line between the two solar D lines very distinctly. 



A NEW FORM OF SPECTROSCOPE 



341 



three component prisms, to find a combination which is practi- 
cally easy to make, and which is capable of satisfying compara- 
tively high requirements with regard to dispersion. 

It will perhaps still be of interest to see how our construction 
is applied to the fluid ethyl cinnamate, which was preferred by 
Wernicke to the strongly absorptive flint glass. I believe that 
the simplest and most advantageous arrangement is that shown 
in Fig. 4, which represents a combination with three members, 
and is one which I have practically carried out. With careful 
regulation of temperature it gives very good images. 




Fig. 4 

ABD is a glass tube, the ends of which are cut off obliquely 
and cemented to the crown glass prisms ADE and BCD, 
The refracting angle of the resulting hollow prism is 120°, that 
of the glass prisms 35° and 63° respectively. A good quality of 
the fluid was obtained from C. F. Kahlbaum in Berlin. 

The optical constants are as follows : 

Ethyl cinnamate «/>= 1.5607, tip — «c=o-0286, ^=19.6, 
Crown-glass «/>=i.5i7o, tip — ;ic = o.oo85, r = 6i.o, 

from which we find that the dispersion from C to F is I2°.i, or 
about the same as that of eight flint glass prisms of 60^. 

It will perhaps be noticed that in both Fig. 3 and Fig. 4 the 
incident ray strikes the first surface of the prism at the same 
angle of about 45^. The reason is that I wished a comparison 
scale reflected from the surface of the prism to be seen in the 
field of view alongside of the spectrum, and took this fact into 
account in computing the combination. The most favorable 
conditions are obtained when the tube containing the scale is 
nearly at right angles to the observing telescope. In cases 
where the comparison scale is not required, it is well to compute 



342 C. PULFRICH 

the combination so that the angle of incidence on the first sur- 
face is less than 45^, as the loss by reflection will then be 
diminished. 

From what has been said above it is evident that the other 
parts of the spectroscope can be constructed without regard to 
the particular form of prism employed. This circumstance 
makes it possible to exchange any given prism for another one 
of greater or less dispersion, without altering any other part of 
the entire apparatus. It is easy to so arrange the prism-holder 
that the prism can be removed by a suitable handle and replaced 
by another, which may be, for example, an ordinary 30^ prism 
of flint glass, and the observer is thus enabled to work quickly 
and conveniently, sometimes with higher and sometimes with 
lower dispersion, according to the requirements of the special 
case in hand. An instrument provided with three prisms, one 
of the ordinary 30° form, one compound with two members, and 
another compound with three members, would probably answer 
even somewhat exacting requirements. 

We will now undertake to illustrate a little more completely 
the special application of the above principles to the construc- 
tion of a spectroscope. Fig. 5 is a side view of the instrument ; 
Fig. 6 shows the course of the rays, and Fig. 7 the arrangement 
of the slit and eyepiece. 

What first strikes the eye in looking at Fig. 5 is the circum- 
stance that the instrument is mounted so that it stands obliquely. 
For laboratory practice it is possible that this position may be 
more convenient than the customary horizontal one ; at least it 
has been found so in the experiments which have been made up 
to the present time. This is, however, merely a matter of taste, 
in which, of course, no fixed standard can be referred to. 

A short description will indicate clearly the relation of the 
different parts. P is the compound prism, fastened to a plate 
which can be rotated about an axis A. The prism is rotated 
through any considerable angle by hand, the angle being read 
on the arc 7*, which is graduated to half degrees by means of an 
index. The slow motion is effected by means of the micrometer 



A NEIV FORM OF SPECTROSCOPE 



343 



screw J/, after the clamp K has been tightened. With this 
arrangement the angle can be read to one division of the microm- 
eter head, or lo'. The parts just described are in a position 
where they are easily reached by the right hand of the observer, 
whose arm rests on the table. The micrometer head and grad- 
uated arc are also read without difficulty. On the left side of 




Fig. 5 {% actual size) 

the observer, opposite to the prism, and perpendicular to the 
telescope, is fixed the tube which contains the scale and its colli- 
mator Oi. The scale is divided on lamp-black from X 750 to X 400 
in units of the second decimal place. By means of a small screw 
E, the tube can be rotated about an axis parallel to that of the 
prism, so that the observer can always adjust the scale to read 
correctly. Between the prism and (?, is a movable screen which 
shuts off the light from the scale when necessary, and intercepts 



344 



C. PVLFRICH 



the reflected images of the slit produced by the two surfaces of 
the objective 0^. 

A particular interest naturally attaches to the arrangement of 
the observing telescope F, which at the same time fulfils the 
function of a collimator. The objective Ob has an aperture of 
2 1 mm j^j^j ^ focal length of 255°*", and the two eyepieces belong- 




FiG. 6 

ing to the instrument magnify 6 and 1 2 times respectively. In 
the focal plane of the objective is placed the slit apparatus shown 
in Fig. 7. The slit can be set to any required width by means 
of the graduated head S, and there is no danger of injuring the 
jaws by turning the screw too far. The slit is so arranged that 
the lower half of the field of view is left entirely free for obser- 
vation. /, and /, are two reflecting prisms, placed alongside of 
each other over the slit plate, which allow the spectra of two 
sources of light, one on the right and one on the left, to be 
simultaneously seen and compared by the observer. The two 



A NEW FORM OF SPECTROSCOPE 345 

spectra 5, and 5, are observed in the lower half of the field with 
the eyepiece. All false light coming from the prisms /, and p, is 
prevented from reaching the eye by a plate which covers the upper 
half of the field of view. The mark used in setting, which is not 
shown in Fig. 7, is a double cross ruled on glass, and must be so 
placed that the two points of intersection are in the line of the 
slit produced, and each separately is nearly in the middle of its 
corresponding spectrum. 

The method for exactly focusing the slit is the same that I 
used in constructing the Abbe-Fizeau dilatometer,' and it is one to 




Fig. 7 (actual size) 

be recommended for any telescope when the principle of reversed 
path is applied. It consists merely in focusing the objective 
instead of the slit. A small pin /, projecting through the tele- 
scope tube, serves as a handle for moving the slide. The great 
advantage of this arrangement is that the sources of light reflected 
by/,/, need be only once adjusted, and do not afterwards require 
any change. The axial displacement of the objective has no 
effect on this adjustment. 

For the study of different sources of light a number of acces- 
sories are provided, which are generally held by the arms H 
when the instrument is in use. They are : a pair of plane mirrors 

« 2r./. Insirum, 1893. p. 376. 



346 C. PULFRICH 

movable in any direction, a pair of ordinary convex image 
lenses, holder for test-tubes, support for absorption cells (the 
last two placed close to the small apertures in front of the reflect- 
ing prisms), and finally a special mounting for ''end-on" Geissler 
tubes. The two lenses, one of which is shown at B^ in Fig. 6, 
and the apparatus B, in the same figure, serve to converge the 
rays from the luminous source upon the slit. The apertures and 
distances of the lenses in the system are so chosen that the 
objective is filled with light. The observer can easily assure 
himself that the source of light is properly adjusted, by first test- 
ing with a piece of paper the course of the rays where they 
enter the small apertures, and then removing the eyepiece, by 
looking in directly at the objective. It will be seen at once what 
part of the objective is covered by the rays and whether any 
further adjustment of the source is necessary. 

In working with sunlight, which is thrown upon the prisms by 
the plane mirrors, it is advisable to produce the desired diver- 
gence of rays passing through the slit by means of diffraction, 
closing the slit until the length of the first diffraction spectrum 
13 just equal to the diameter of the objective. In this operation 
the method of viewing the objective which has already been 
described is particularly useful. It is only necessary to observe 
that the middle and brightest band of the large number of bright 
and dark bands which are produced when the slit is slightly 
w.dened, falls centrally upon the objective. If then the slit is 
narrowed until only the central band remains, the eyepiece can 
be replaced, and the Fraunhofer lines must then appear with 
maximum distinctness. The desired divergence of the solar rays 
can also be produced by employing a concave instead of a plane 
mirror, but, as it seems to me, in a less advantageous manner. 

I have still to describe the manner in which the annoying 
reflections from the two surfaces of the objective are avoided. 
The simp' est and most effective way is to stop the reflections by 
suitable screens placed at the points where the images are 
formed, but this of course is only possible when the images are 
real. Whether they are real or not, and where they are situated. 



A NEW FORM OF SPECTROSCOPE 347 

depends upon the special construction of the objective, and par- 
ticularly upon the relations between the curvature of the 
surfaces. 

The objective of the instrument we are considering has a flat 
outer surface and a convex inner one. The image of the slit 
formed by the flat surface is in the plane of the slit, and by 
a minute inclination of the objective it can be made to fall upon 
the back of the slit-plate, so that with respect to any effect on 
the observations it is practically non-existent. Instead of inclin- 
ing the whole objective it is advisable to place only the outer 
(plano-convex) lens slightly oblique to the optical axis, and this 
can be done when the two lenses are cemented together. An 
extremely minute prismatic displacement of the spectrum is thus 
produced, but as it is in the direction of the lines it has no effect 
whatever upon their appearance. 

The other reflected image of the slit which it is necessary to 
consider is produced by the convex surface of the objective. It 
is virtual, is situated immediately behind the objective, and can- 
not therefore be readily suppressed. The effect of this second 
image, which is seen at once as a bright point of light when the 
objective is viewed directly, is to produce a slight, uniform illu- 
mination of the field of view. The illumination is faint, because 
ther rays from the image are strongly divergent. This difficulty 
can also be remedied, either by placing a very small screen out- 
side the eyepiece (at the position of the eye) to intercept the 
image, or, what appears to me to be the simpler way, by placing 
a drop of black varnish on the center of the objective. The lat- 
ter arrangement does not have the least injurious effect on the 
observation of the spectral lines, and the loss of light which it 
causes is imperceptible. 

With our apparatus, therefore, the difficulties arising from 
the reflection of light by the surfaces of the object glass are 
completely overcome. 

All other false light is prevented from reaching the eye by 
the eyepiece cap Oc^ a protecting screen supported by the eye 
end of the telescope, and a cover which goes over the prism and 



348 



C. PULFRICH 



the two objectives Ob and 0^. The last two protective devices 
are not shown in Figs. 5 and 6. 

The application of the spectroscope described above is 
restricted to qualitative investigations. In order to adapt the 
instrument to wider fields of research (photometry, etc.) it is 
necessary to separate the collimator from the observing tele- 
scope. This can be done by placing an ordinary totally reflect- 
ing prism in front of the objective of the collimator tube, which 
is fixed at right angles to the telescope, so that the rays reflected 
from it are made parallel to the axis of the latter. The two 
tubes must now of course be placed at different heights (the 




Fig. 8 

distance between centers being equal to the diameter of the 
objectives). The prism /^must consequently have a height equal 
to twice the diameter of the objectives, and the rays are trans- 
ferred from one story to the other by two total reflections at the 
interior surfaces of the prism R (Fig. 8), which is cemented to 
the back of the large prism. This form of instrument retains in 
other respects all the advantages which have previously been 
described. 

An article by A. E. Tutton on "An Instrument of Precision 
for producing Monochromatic Light of any desired Wave-^length, 
and its Use in the Investigation of the Optical Properties of Crys- 
tals"' gives me occasion to point out that our apparatus (in both 
forms) can also be applied to the purposes which are there men- 

«/Vwf. ^. 5. 55, III, 1894. 



A NEW FORM OF SPECTROSCOPE 349 

tioned, and, it seems to me, much more advantageously. It is only 
necessary for this purpose to cover the lower half of the field of 
our spectroscope with a plate, having a slit parallel to the one 
already in place. Such a second slit, if movable, would also be 
of great value for visual observations, as it would enable the 
observer to examine any given part of the spectrum while all 
other parts were excluded. 



Minor Contributions and Notes. 



PHOTOGRAPHIC CORRECTING LENS FOR VISUAL 
TELESCOPES. 

Since my article on a correcting lens was printed in the February 
number of The Astrophysical Journal I have learned a number of 
interesting facts relating to the history of the subject. It seems that a 
lens of the form considered in my article was described by the Astron- 
omer Royal (Mr. Christie) at a meeting of the Royal Astronomical 
Society on June lo, 1887; but as the lens was intended chiefly for 
ordinary photography of small objects, Mr. Christie withdrew his 
paper, with the intention of working out a combination which would 
give a larger field. Hence the paper was not published. Dr. Huggins 
saw at once, however, that the lens was just the thing for photographic 
work with the spectroscope, and at his request Mr. Turner computed 
a lens according to Mr. Christie's plan from data which Dr. Huggins 
furnished. The lens was made by Sir H. Grubb. Its diameter is three 
inches, and it is placed twenty-seven inches within the focus of the 
fifteen-inch refractor. The alteration of focus for the F line is only 
about one-quarter of an inch, which is of course immaterial, and the 
spectrum is practically linear from F to beyond H. It is only quite 
recently that this lens has been used for spectrum photography. 

The discussion at the meeting of the Royal Astronomical Society is 
reported in the Observatory^ lo, 2 5 5, but as the subjects of such discussions 
are not indexed, it naturally escaped my attention. The theory of a 
correcting lens for axial pencils is so simple that I should have hesitated 
to print my article if I had seen even this passage in the Observatory; still 
I think the curves which I have computed are interesting, as showing what 
may be expected from such a lens in the case of a very large telescope. 

Mr. Newairs single corrector shortens the focus of the Cambridge 
twenty-five- inch refractor by only about nine inches, and the convergence 
of the refracted cone is i : 10. All the observations, both visual and pho- 
tographic, are made in the altered focal plane. This arrangement would 
be impracticable in the case of a very large refractor, as with a ratio of 
1 : 10 the focus would be far within the tube. James E Keeler 

350 



MINOR CONTRIBUTIONS AND NOTES 35 1 

THE COLOR OF SIRIUS IN ANCIENT TIMES. 

In a paper on "The History of the Color of Sirius," which appears 
in Astronomy and Astro-Physics, Vol. XI (for 1892) p. 376, Dr. T. J. J. See 
refers to Virgil's mention of the star in the Georgics, but seems to have 
overlooked two similar places in the ^neid. Perhaps you will allow 
me to supply the omission, though it does not materially affect the 
conclusion drawn. 

The first passage is in the third book of the /Eneid, v. 141, and runs 
thus: "tum steriles exurere Sirius agros," which certainly seems to 
suggest the idea of red flames. 

The other is in the tenth book, where we read (vs. 272, 273): 

" Non secus, ac liquida si quando nocte cometse 

Sanguinei lugubre rubent, aut Sirius ardor.** 

Sirius is here taken as an adjective, but the sense is clear ; the com- 
parison to the reddening light of a comet presaging slaughter shows 
that '' ardor '^ conveys the idea of red light, similarly to the passage 
" ardebat in coelo," quoted by Dr. See from the fourth book of the 
G'orgics. W. T. Lynn. 

Blackheath, London, S. E. England, Feb. 18, 1895. 



ON THE VARIABILITY OF ES.BIRM, 281. 

This star is Schj. 115 ; Birm, 211 ; BD. + 17°. 1973 ; R. A. 8** 
49" 45' ; Dec. N. 17° 36'. 7 (1900). 

Schj. note is "Cape obs.: fine red; brick red; 8.5. T. Mayer, 
380; 7. LI. 17,576; 6.5. Chacornac; 6. Birmingham; 7 bis 7.5.** 
Birmingham says: "Several since those quoted by Schj. from A. N,, 
1843 ; color varying in different degrees of red and orange ; 7-7.5." 
It is No. 69 of a Catalogue of " Suspected Variable Stars," communi- 
cated by W. H. St. Q. Gage, T. Read and myself to the English 
Mechanic, June 7, 1882, and subsequent numbers. The note there is 
6^-8>^ var. color Birmingham. It was copied by Mr. Gore into his 
" Catalogue of Suspected Variable Stars" (Proc. Roy. Irish Acad. 4, 
No. 3) and is No. 274. The variation in color mentioned by Bir- 
mingham is dropped out and the extremes of magnitude retained. I 
noted in the new edition of Birmingham " the star is probably a vari- 
able of the 19 Piscium type." As far as I am aware the only observations 
from comparison stars are : 



352 MINOR CONTRIBUTIONS AND NOTES 



1878, Feb. 


28, 7.3, 


Gore. 


1885, Feb. 


J, 6.4, 


Es. 


1878, Dec. 


21, 7>^» 


« 


1885, " 


9. 6-8. 


« 


1883, Nov. 


22, 7J4, 


« 


1885, " 


15. 6-7. 


Gage. 


1884, April 


11, ±:7» 


K 


1885, " 


16, f>^. 


Gore. 


1884, Dec. 


20, 6.5, 


GaRC. 


1885, " 


17. 6.5, 


£s. 


1884, " 


21. 6.8, 


« 


1885, Mar. 


'''N.7 
14. J 


riofTi 


1884, " 


22, 7.0, 


<( 


1885, April 


1 vj*5< 


1885, Jan. 


3. 6.5, 


« 


1885, Dec. 


16, 6.5, 


£s. 


1885, " 


21, 6.3, 


« 


1887, J^- 


29. 6.5, 


« 



Since then, apparently, it has not been observed. Gage's observa- 
tions and mine should be 1884-5, °ot 1881-2 as in Es.-Birm. The 
following stars are suitable for comparison stars : 

A BD. + 17°. 1979, I" 46V; 5' ^- Mag. 6.8. 
B BD. + I7°.i966, 2" lo*/; 8' n. Mag. 7.7. 
In all my previous observations the star had been rather brighter 
than A, save 1882, February 9. On looking the star up with a bin- 
ocular on January 26, 1 was much astonished at its faintnef^s. This 
night was poor and the stars unsteady ; it was reobserved on January 
28, 29. The following are the results : 

1895, Jan. 26, ± 7.2. 
1895, Jan. 28, 7.8. 
1895, Jan. 29, 7.6. 
These observations compared with the previous ones of mine seem to 
leave no doubt as to its variability. It is of type IV. j £ Espin 
Tow Law, Darlington, January 30, 1895. 



THE DISPLACEMENT OF SPECTRAL LINES CAUSED BY 
THE ROTATION OF A PLANET. 

M. Deslandres has recently published a note (C R. zao, 417- 
420) on the application of Doppler's principle to the determination of 
the period of rotation of a planet, showing that in the case of a rotat- 
ing body like Jupiter,' which shines by reflected light, the displace- 
ment of a line in the spectrum is twice as great as it would be if the 
body were self-luminous. The theory of this supposed new principle 
is discussed by M. Poincar^. 

The principle has however long been recognized, although I am 
unable to say definitely by whom it was first stated. Perhaps the 

> Supposed to be in opposition. 



MINOR CONTRIBUTIONS AND NOTES 353 

earliest reference to the effect of the motion of a reflector is by Niven 
(J/. N, 34, 345, 1874), who calculated the relative motion of the Earth 
and each of the other planets in the line of sight on the assumption of 
circular motion. The motion of a planet in the line of the radius 
vector was therefore neglected, although its effect was pointed out. 

Mr. Maunder has clearly described the effect of rotation in this 
very case of Jupiter. He says: "The planet Jupiter, from the won- 
derful velocity with which it rotates, affords another means of 
demonstrating this displacement to the eye. The difference of motion 
of the two limbs is more than 15 miles per second, and the effect of 
this difference is doubled by the fact that Jupiter shines by reflected, 
not inherent light. The relative displacement therefore is equiva- 
lent to that produced by a motion of 30 miles per second, a very appre- 
ciable amount" {Obsy.%, 118, 1885). 

In some photographs of the spectrum of Jupiter that I made 
last winter with high dispersion, the obliquity of the planetary lines 
is shown very beautifully. The planet was kept very accurately cen- 
tered, with the slit parallel to the belts, and a solar spectrum was 
subsequently photographed on each side. I made no careful meas- 
urement of the inclination of the lines, as the plates were made for 
another purpose, but satisfied myself that it represented a difference 
of velocity of about 30 miles. James E. Keeler. 



Dr, Pulfrich^s Modification of the lAttrow Spectroscope, — Although 
the spectroscope described in the paper by Dr. Pulfrich (translated in 
our present number) has little of novelty beyond the substitution of a 
compound half-prism for the simple half-prism of the ordinary form 
of Littrow spectroscope, it is carefully worked out, and is put into the 
practical and convenient form which is characteristic of the Zeiss work- 
shops. For laboratory purposes requiring only a moderate dispersion 
it should be a useful instrument. We do not share the author's opinion 
as to its advantages (other than those of compactness) when a con- 
siderable dispersion is given to the apparatus by increasing the num- 
ber of pieces in the compound half-prism. It would then have the 
same defects that make the ordinary direct-vision prism unserviceable 
for work requiring brightness of spectrum and good definition, namely, 
great thickness of glass to be traversed and comparatively small resolv- 
ing power. For such work nothing can replace the prism-train of the 
usual form of spectroscope. 



Recent Publications. 



A LIST of the titles of recent publications on astrophysical and 
allied subjects will be printed in each number of The Astrophysical 
Journal. In order that these bibliographies may be as complete as 
possible, authors are requested to send copies of their papers to both 
Editors. 

For convenience of reference, the titles are classified in thirteen 
sections. 

1. The Sun. 

Burton- Brown, Col. A. Notes on the coming Toul Eclipse of the 
Sun, August 8, 1896, and Suitable Observing Stations in Norway 
M. N. 55, 93-101, 1895. ]ovLt, B. A. A. 5, 158-163, 1895. 

Deslandres, H. On the Electric Origin of the Solar Chromosphere. 
Knowl. x8, 59-60, 1895. 

GuiLLAUME, J. Observations du Soleil, faites i, I'Observatoire de 
Lyon (Equatorial Brunner) pendant le quatridme trimestre de 1894. 
C. R. xao, 250, 1895. 

LocKYER, J. Norman. The Sun's Place in Nature. Nat. 51, 374-377. 
1895. Nat. 51, 396-399, 1895. 

MONCK, W. H. S. The Sun's Motion in Space. Pub. A. S. P. 7, 33- 
38. 1895. 

Stratonoff, W. Bestimmung der Rotationsbewegung der Sonne aus 
Fackelpositionen. A. N. X37, 165-167, 1895. 

Sykora, I. Osservazioni spettroscopische solari fatte a Charkow. dur- 
ante il 1894. Mem. Spettr. Ital. 23, 201-207, December, 1894. 

Townsend, J. S. Observations of Solar Prominences. Jour. B. A. A. 

5, I53» 1895. 

White, W. Hale. The Wilsonian Theory of Sun-Spote. Jour. B.A.A. 
5, 218. 1895. 

. Mean areas and heliographic latitudes of Sun-Spots in 

the year 1892, deduced from photographs taken at the Royal Observ- 
atory, Greenwich, at Dehra DOn (India) and in Mauritius. M. N. 
55i 150-154, 1895. 

2. The Solar Spectrum, Infra-Red, Visible, and Ultra-Violet. 

LocKYER, J. Norman. Observations of Sun-Spot Spectra 1879- 1894. 
Proc. R. S. 57, 199, 1895. 

354 



RECENT PUBLIC A TIONS 355 

3. Stars and Stellar Photometry. 

Campbell, W. W. . The New Star of 1892. Pub. A. S. P. 7, 31-33 

1895. 
OuDBMANS, J. A. C. Ueber die Aenderung der Helligkeit der Fixsterne 

Zufolge der eigenen Bewegung in der Richtung der Gesichtslinie. 

A. N. X37, 169-171, 1895. 
Parkhurst, H. M. Notes on Variable Star8.-No. 7. A. J. No. 339, 15, 

20-22, 1895. 
Parkhurst, J. A. Observations of Suspected Variables. A. J. No. 339, 

X5, 24, 1895. 
Peek, Cuthbert E. Observations of Variable Stars. Jour. B. A. A. 

5, 212-214, 1895. 
Radclippe Obs*y, Oxford. Estimations of Magnitude of Nova Auri- 
ga. M. N. 55, 164, 1895. 
Rambaut, a. a. On the Effect of Atmospheric Refraction on the 

Position of a Star. M. N. 55, 123-145, 1895. 
Sawyer, Edwin F. Observations of Variable Stars. A. J. No. 338, 15, 

11-13. 1895. 
ScHJBBBRLE, J. M. Note on the Visual Magnitude of the Faintest Stars 

photographed by Dr. Max Wolf, at Heidelberg. A. J. No. 338, 15, 

16, 1895. 
Turner, H. H. Note on Professor E. C. Pickering's " Comparison of 

Photometric Magnitudes of the Stars.** A. N. 137, 155-156, 1895. 
Turner, H. H. On the Measurement and Reduction of the Plates for 

the Astrographic Chart. M. N. 55, 102-118, 1895. 
Yendell, p. S. Ephemeris of Variables of the Algol-Type. A. J. No. 

339, 15, 18-20. 1895. 

4. Stellar Spectra, Displacements op Lines and Motions in the 

Line op Sight. 

LocKYER, J. Norman. On the Photographic Spectrum of 7 Cassiopeiae. 

Proc. R. S. 57, i73-»77. 1895. 
MoNCK, W. H. S. The Spectra and Proper Motions of Southern Stars. 

lour. B. A. A. 5, 164, 1895. 

5. Planets, Satellites and their Spectra. 

Cammell, B. E. Observations of Mars. Jour. B. A. A. 5, 1 48-1 51, 

1895. 
Campbell, W. W. The Irregular Waning of the South Polar Cap of 

Mars. Pub. A. S. P. 7, 40-43, 1895. 
Denning, W. F. Jupiter. Nat. 51, 227-229, 1895. 



356 RECENT PUBLIC A TJONS 

Flam MARION, C Le monde g^ant de Jupiter. Bull. Mens. Soc. Astr. 

France. 30-37, February, 1895. 
Henderson, A. Observations of Jupiters* Red-Spot. Jour. B. A. A. 5, 

154, 1895. 
Klein, J. Welches sind die Dimensionen des kleinsten auf dem Mond- 

photographien sichtbaren Details? Sirius, as, 26-32, 1895. 
Landerer. J.-J. Sur un passage de Tombre du quatridme satellite 

de Jupiter. C. R. Z20, 248, 1895. 
Manson, Marsden. The Climate of Mars. Pub. A. S. P. 7, 53-57, 

1895. 
Maunder, E. Walter. The "Eye" of Mars. Knowl. x8, 55-59, 1895. 
Pater SON, A. G. Observations of Jupiter's Red-Spot. Jour. B. A. A. 

5, 211-212,1895. 
Roberts, C. An Observation of Saturn. Jour. B. A. A. 5, 2 19-220, 1895 . 

7. NEBULiE AND THEIR SPECTRA. 

Wolf, Max. Notiz Uber die Plejaden-Nebel. A. N. 137, 175, 1895. 

8. Terrestrial Physics. 

Douglass, A. E. The Study of Atmospheric Currents by the Aid of 
Large Telescopes, and the Effect of Such Currents on the Quality 
of the Seeing. Am. Met. Jour. 11, 395-412, 1895. 

Gem MILL, S. M. B. The Zodiacal Light. Jour. B. A. A. 5, 216-217, 
1895. 

Herschel. a. S. Aurora of November 23, 1894. Nat. 51, 246-247, 

1895. 
. Aurora of December 23, 1894. Jour. B. A. A. 5, 52, 

1895. 
Perry, John. On the Age of the Earth. Nat. 5X1 341-342, 1895. 

9. Experimental and Theoretical Physics. 

BoLTZMANN, LuDwiG. On Certain Questions of the Theory of Gases. 

Nat. 51, 4I3-4I4» 1895. 
EuMORFOPOULOS, N. On the Determination of Thermal Conductivity 

and Emissivity. Phil. Mag. 39, 280-294, March, 1895. 
Jaumann, G. Bemerkung zu der Abhandlung Uber Lichtemisaion. 

Wied. Ann. No. i, 54, 178-180. 
Knoblauch, Osc. Ueber die Fluorescenz von LOsungen. Wied. Ann. 

No. 2, 54, 193-220, 1895. 
Moreau. G. Sur la dispersion rotatoire anormale des milieux abior* 

bants cristallis^s. C. R. xao, 258, 1895. 



RECENT PUBLIC A TIONS 357 

Nichols, Edward L. and Mary C. Spencer. The Influence of Tem- 
perature upon the Transparency of Solutions. Phys. Rev. a, 344- 
360, 1895. 

Porter, Alfred W. On the Influence of the Dimensions of a Body 
on the Thermal Emission from its Surface. Phil. Mag. 39, 268-279, 
March, 1895. 

Sav^lief, R. Sur le degr^ de precision que Ton peut atteindre dans 
les observations actinom^triques. Ann. Chim. et Phys. 4 (7), 424- 
429, 1895. 

Schuster, Arthur. The Kinetic Theory of Gases. Nat. 51, 293, 1895. 

Thielb, E. Spektrophotometrische Untersuchung der verschiedenfar- 
bigen Jodldsungen. Zeit. Phys. Chem. No. i, 16, 147-155, 1895. 

Warburg, E. Ueber W^rmeleitung und Temperatur der in Geissler*- 
schen R5hren leuchtenden Case. Wied. Ann. No. 2, 54, 265-275, 
1895. 

11. Photography. 

Baldcock, J. H. The Influence of Orthochromatic Plates and Yellow 
Screens on Colour Values. Brit. Jour. Phot. 42, 70, February i, 1895. 

Janssen, J. La photomdtrie photographique. Annuaire du Bureau 
des Longitudes, C. i, 1895. 

12. Instruments and Apparatus. 

Borgesius, a. H. Beschreibung eines Interferenzrefractometers, 

Molecularrefraction und Dispersion einiger Sake in Losungen. 

Wied. Ann. No. 2, 54, 221-243, 1895. 
Common, A. A. Preliminary note on a modified (oblique) form of Cas- 

segrain telescope. M. N. 55, 86-88, 1895. 
Houston, E. J. and A. E. Kennelly. A New Method of Measuring 

Illumination. Elec. Eng. 19, 226, March 6, 1895. 

13. General Articles, Memoirs and Serial Publications. 

Gladstone, J. H. Argon. Nat. 51, 389-390, 1895. 

von KONKOLY, N. Beobachtungen angestellt am Astrophysikalischen 

Observatorium in O Gyalla. H. W. Schmidt, Halle 1894, 15-16, 

107 pp., 1892-3. 
VoGEL, H. W. Die farbigen Wasser der Caprenser Grotten, der 

Schweizer Eishdhlen und Yellowstonequellen. Wied. Ann. No. i, 

54, 175-177. 1895. 



NOTICE. 

The scope of The Astrophysical Journal includes all investigations of 
radiant energy, whether conducted in the observatory or in the laboratory. 
The subjects to which special attention will be given are photographic and 
visual observations of the heavenly bodies (other than those pertaining to 
"astronomy of position'*); spectroscopic, photometric, bolometric and radio- 
metric work of all kinds ; descriptions of instruments and apparatus used in 
such investigations ; and theoretical papers bearing on any of these subjects. 

In the department of Minor Contributions and Notes subjects may be 
discussed which belong to other closely related fields of investigation. 

It is intended to publish in each number a bibliography of astrophysics, 
in which will be found the titles of recently published astrophysical and 
spectroscopic papers. In order that this list may be as complete as possible, 
and that current work in astrophysics may receive appropriate notice in other 
departments of the Journal, authors are requested to send copies of all 
papers on these and closely allied subjects to both Editors. 

Articles written in any language will be accepted for publication, but 
unless a wish to the contrary is expressed by the author, they will be translated 
into English. If a request is sent with the manuscript twenty-five reprint 
copies of each paper, bound in covers, will be furnished free of charge to the 
author. Additional copies may be obtained at cost price. No reprints can 
be sent unless a request for them is received before the Journal goes to 
press. 

The Editors do not hold themselves responsible for opinions expressed by 
contributors. 

The Astrophysical Journal will be published on the first day of every 
month except July and September. The annual subscription price for the 
United States, Canada and Mexico is $4.00 ; for other countries in the Postal 
Union it is 18 shillings. Correspondence relating to subscriptions and adver- 
tisements should be addressed to The University of Chicago, University Press 
Division, Chicago, III, 

IVm, Wesley &* Son, 28 Essex St., Strand, London, are sole foreign agents, 
and to them all European subscriptions should be addressed. 

All papers for publication and correspondence relating to contributions 
and exchanges should be addressed to George E, Hale, Kenwood Observatory, 
Chicago, III, 



358 



> 

X 



< 

l-J 
a. 












\ 



O 

H 
< 

> 

c/) 

PQ 
O 

Q 
O 
O 

z 

:>< 

H 



X 

a* 
< 

O 

o 

H 

o 

X 

a* 



u 

Z 

o 

a: 
a* 

> 

H 

a* 

D 



U 

< 



THE 

ASTROPHYSICAL JOURNAL 

AN INTERNATIONAL REVIEW OF SPECTROSCOPY 
AND ASTRONOMICAL PHYSICS 



VOLUME I 



MAY 1895 NUMBERS 



THE MODERN SPECTROSCOPE. XIL 

THE TULSE HILL ULTRA-VIOLET SPECTROSCOPE. 

By William Huggins. 

The spectroscope which I designed and had constructed 
during the seventies for my original work on photographing the 
spectra of the stars,' was arranged to include the whole of the 
ultra-violet region of the light from the heavenly bodies which 
reaches the Earth. 

I had at my command a refractor of fifteen inches aperture, 
and a Cassegrain telescope of eighteen inches aperture, both 
belonging to the Royal Society. I chose the latter instrument 
for my work, notwithstanding the drawback of some want of 
permanency of the collimation of the mirrors, on account of the 
freedom of a reflector from the outstanding chromatic aberra- 
tions of a refractor, and also because by the employment of 
Iceland spar for the prism and quartz for the lenses, the whole of 
the more refrangible part of the spectrum could be photographed, 
at least as far as the absorption of our atmosphere allows rays of 
small wave-length to pass. 

' "On the Photographic Spectra of the Stars." Phil. Trans, 1880, Part ii, p. 669. 

359 



360 WILUAM MUGGINS 

The Cassegrain telescope, which has mirrors of speculum 
metal of very fine defining power, was made by Sir H. Grubb. 
In 1882 it was mounted, by the novel device of a double declina- 
tion axis, one axis moving within the other, as a twin telescope, 
together with the fifteen-inch refractor, upon the same equatorial 
stand. This instrument has been used chiefly for spectroscopic 
work, but last year advantage was taken of the fine definition of 
the specula to make some crucial observations of the character 
of the image of Nova Aurigae.' 

The early arrangement employed in 1876-1879 consisted 
essentially of a small spectroscope containing a single prism of 
Iceland spar, and lenses of quartz, the slit of which was placed 
in the principal focus of the great speculum, eighteen inches in 
diameter, of the Cassegrain telescope, the small convex speculum 
having been removed. 

In this instrument the plan was adopted for the first time of 
turning the jaw plates of the slit into mirrors, in which the 
objects to which the instrument was directed could be seen by 
reflection at the same time as the slit itself. In the first instance 
polished silver was used for the reflecting substance ; afterwards 
very thin plates of quartz, silvered at the back, the edges of which 
formed the slit ; and finally in the new spectroscope attached to 
the refractor.' speculum metal was found to fulfil very satisfac- 
torily the necessary conditions of giving a permanently reflect- 
ing surface, and furnishing true edges for the slit. In this early 
instrument the images of celestial objects reflected from the 
mirror-jaw plates were observed through the hole in the great 
speculum by means of a small telescope fixed in the place of the 
eyepiece. 

The advantage which this form of spectroscopic arrangement 
possessed of reducing the loss of light by reflection to that at the 
surface of one speculum only, was accompanied by some draw- 

'In this instrument, which is of course free from chromatic aberration, the 
images of Nova Aurigse and of the star near it were indistinguishable in character 
under a magnifying power of 700 diameters. Both appeared equally stellar. A, *N, 
321 1. A, and A, April, 1894, p. 314. 

* For photograph and description see Astronomy and Asiro-Phyms, 



THE. MODERN SPECTROSCOPE 36 1 

backs. The spectroscope, though made as small as possible, 
was larger than the four-inch hole in the speculum, and blocked 
out some light. The adjustments of the spectroscope itself, and 
also of its relation to the speculum, could only be made with some 
inconvenience at the top of the tube. For the same reason, 
unless the telescope was directed to an object very low down, it 
was necesssLry, at some loss of time, to unclamp it in declination 
and bring the spectroscope-end within reach in order to insert or 
to change the photographic plate. 

There was the further disadvantage that in consequence of 
the large ratio of aperture to focal length of the great speculum, 

namely, -^» the collimator had to be made very short. 

Consequently with one prism, to which the spectroscope was 
necessarily restricted on account of size, either light or the 
necessary purity of the spectrum had to be sacrificed. If the 
slit were opened wide enough to just include, or even nearly so, 
the image of a star, its angular magnitude relatively to the dis- 
persion was too great for the needful resolution of lines, and 
therefore, as a matter of fact, the slit was always used too narrow 
to receive more than a part of the light of a star, with the great 
disadvantage of long exposures. 

The new instrument is free from these disadvantages, though 
in one respect it comes short of the earlier arrangement, since 
there is additional loss of light from reflection at the second 
speculum. The Cassegrain telescope is restored to its original 
form, and the collimator of the new spectroscope passing up 
through the hole in the large speculum, the slit is placed within 
the telescope tube at the focal plane after reflection from the 
small convex speculum. 

In Fig. I, Plate XVI, the collimator is seen within the tele- 
scope tube ; in Fig. 2, Plate XVII, the remaining part of the 
spectroscope, outside and below the telescope tube, is shown. 

Returning to Fig. I, the diagram explains itself. The slit is 
adjusted by means of a rod, which in Fig. 2 is seen to pass 
below the spectroscope and to terminate in a graduated head. 



362 WILLIAM MUGGINS 

Behind the slit slides a small shutter which closes the central 
half of the slit, to protect the part of the plate on which the 
star's spectrum falls, when, either before or after exposure, 
narrow comparison spectra are photographed through the outer 
parts of the slit, above and below the star's spectrum. 

In front of the slit extends a tube twelve inches long, 
furnished at the end with a sliding diaphragm having an opening 
of such a size as to exclude all light except that reflected from 
the small convex mirror. 

A very successful arrangement of the slit-mirror method has 
been adopted by which the slit, together with a small field of 
stars, can be conveniently seen by an observer looking into the 
diagonal eyepiece, shown in Fig. 2, This eyepiece, by means 
of the clamp, can be brought into and then fixed in the position 
which is most convenient for observation. The polished slit- 
plates of speculum metal are slightly inclined so that the light 
which does not pass on through the slit is reflected, as shown in 
Fig. I, to one side of the diaphragm-tube. There it falls upon 
the first face of a prism of such a form that after two internal 
reflections it returns along the small view-tube placed by the side 
of the collimator. A few inches below the second surface of the 
prism is placed a small achromatic lens having a focal length 
equal to its distance from the slit. The rays after passing through 
it are rendered parallel, and then pass on without loss, until at a 
little distance from the eyepiece. Fig. 2, they encounter a second 
achromatic lens. This has a focal length of about six inches, and 
with a suitable eyepiece gives a well-defined and bright view of 
the small field of stars upon the slit-plates. On a dark night, or 
when an object of finite magnitude, as a planet or a nebula, is 
not upon the slit, the opening of the slit becomes lost to view. 
For the purpose, under these circumstances, of illuminating the 
slit artificially, a very small incandescent lamp made of ruby 
glass is inserted through the side of the diaphragm-tube a little 
way from the slit. Fig. i. It is enclosed so that light passes 
only upon the slit-jaws. From the position of the lamp it will 
be seen that its light is not reflected back from the slit in the 



^:^ 






1 li 






u:i 



> 

X 

u 



J k 



I:? I 



§ 

o 

H 

u 
u 
a< 

CO 

H 
O 



H 

»-4 



2 






O 



Z 

UJ 



> 

H 
Pi 




THE MODERN SPECTROSCOPE 3^3 

direction to pass to the observer's eye. The slit*jaws are 
illuminated by that small part only of the red light of the lamp 
which is scattered from the mirrors in consequence of imperfect 
polish. This feeble illumination is found in practice to be just 
what is needed to show the slit distinctly, without overpowering 
faint objects. A variable resistance is placed within reach, so as 
to make it easy to obtain with exactness the precise degree of 
illumination which is most suitable to each object. From the 
position of the lamp any light which passes through the slit does 
not pass on to the collimator-lens, but is absorbed by the 
blackened inner surface of the tube. The ease with which the 
slit can, by this arrangement, be placed with precision upon a 
star, or upon a small part of a planet or of a nebula, is all that 
can be desired. 

In Fig. I the detached tube terminated by a right-angled 
reflecting prism forms part of the arrangement for throwing into 
the slit the light from sparks or flames for comparison. When 
in use this tube, which slides through an outer tube furnished 
with the necessary adjusting screws fixed upon the outside of 
the telescope*tube, is pushed in until the reflecting prism comes 
in front of the opening at the end of the diaphragm-tube before 
the slit. The light from the spark, vacuum tube, or flame out- 
side the great telescope passes through a double quartz-com- 
bination fixed in the tube near the outer end, by which it is made 
to converge to a focus upon the slit, and then to diverge at a 
little greater angle than is necessary to completely fill the coUi- 
mator-lens. 

When not in use, this tube can be wholly withdrawn outside 
the telescope^tube, so as not to intercept any light from the great 
mirror. 

Fig. 3 shows the prism box, which contains two prisms of 
Iceland spar, each with a refracting angle of 60^. These were 
made for me by Mr. Hilger and have been cut very successfully. 
The smaller prism, which limits the beam that can pass through 
them, has a length of a^ inches with a height of i^ inches. 

It was decided to work with the prisms in a fixed position. 



364 WILLIAM MUGGINS 

though this position can be varied from time to time for differ- 
ent parts of the spectrum. The prisms and camera are there- 
fore provided with clamps, by which, when all the necessary 
adjustments of the prism, of the camera-lens and of the plate- 
holder have been made, the whole apparatus can be secured 
rigidly in position. All the different adjustments are provided 
with divided scales, so that if it were necessary for any reason to 
dismount any part of it, the instrument could be put back again 
into its former position with great exactness. 

The instrument is provided with two camera- lenses, one of 
about 5 ^ inches, which is now in use for nebulas, and a lens of 9 
inches for stellar spectra, the larger scale of the spectrum mak- 
ing it more independent of the granulation of the gelatine film. 

A range of spectrum from F to a little beyond S, which with 
the shorter camera-lens measures about ^ inch in the plate, can 
be obtained with good definition throughout, and also quite free 
from any duplication due to the double refraction of the spar. The 
spectrum indeed extends some distance beyond S, but after this 
point there is a little falling off in definition. The diameter of 
the collimator lens is i^ inches, that of the camera lens being 
slightly greater. 

For greater lightness the camera*-box is made of aluminium, 
and as the slit is placed in the direction of the star's motion, 
this box stands up in a nearly vertical position when the tele- 
scope is in the meridian, which is a very favorable one for free- 
dom from flexure. 

The spectroscope, as a whole, is secured by means of strong 
clamps within the large tube which screws on to a plate fixed on 
the base of the telescope-tube behind the great speculum. It can 
therefore be attached and removed from the telescope without 
any derangement of its internal adjustments. 

Through the strong supporting tube the spectroscope, as a 
whole, can slide a few inches for the purpose of a first rou£^ 
adjustment of the slit to the focal plane ; the final adjustment is 
then made by a fine screw which gives a slow motion to the small 
convex speculum. 



THE MODERN SPECTROSCOPE 365 

The necessary breadth can be given to the spectra of stars in 
two ways, either by allowing the star's image to trail in the slit, 
or by means of a concavo-cylindrical lens of quartz which is 
mounted in a short tube, which can take the place of the sliding 
diaphragm-tube at the end of the tube before the slit. 

The long equivalent focal length of the Cassegrain form of 
telescope is of advantage in many cases of modern astronomical 
spectroscopy, where it is desirable to have images of consider- 
able dimensions upon the slit-mirrors. It will become, doubtless, 
of increasing importance to be able to photograph separately the 
spectra of adjacent parts of the surfaces of nebulas, and of the 
planets, and to obtain, without enlargement, sufficient breadth of 
spectrum in the case of very small nebulae. Further it will be 
desirable to bring separately upon the slit, and to maintain there, 
the components of binary and multiple stars, and also the stars 
involved in nebulae. The Cassegrain form furnishes the means 
of conveniently obtaining a long equivalent focal length, while 
the instrument itself, and the building covering it, remain of 
moderate dimensions. 



ON THE SPECTROGRAPHIC PERFORMANCE OF THE 
THIRTY-INCH PULKOWA REFRACTOR. 

By A. B^LOPOLSKY. 

When a spectrograph similar to that of the Potsdam Observa- 
tory was ordered for this observatory, in 1890-91, it was decided 
to attach it to the fifteen-inch refractor for the purpose of under- 
taking, as time permitted, preliminary researches in the domain 
of celestial spectroscopy. After the spectrograph had arrived 
here I suddenly received (November 9/21, 1 891) an order from 
the Director " to employ all means, without sparing money or 
time, to adapt the spectrograph as quickly as possible to the great 
thirty-inch refractor." 

This by no means easy task could of course be only partially 
carried out at the time, but it was nevertheless interesting to make 
a test of the instrument, although careful reflection would lead a 
priori to the conclusion that no particularly fruitful results could 
be obtained, for various reasons, among which are these: — the 
objective is corrected only for the visual rays, the dimensions of 
the spectrograph were not designed for the large refractor, han- 
dles for the slow-motions as well as an adapter were lacking, pro- 
vision was not made for observing steps, and, finally, the dome 
was too small in many positions of the instrument when the spec- 
trograph was attached. 

After the most indispensable, but still inadequate, accessories 
were ready, toward the middle of the summer of 1892, my obser- 
vations were begun and have been continued in the summer 
months of 1893 and 1894. 

I wish now to communicate some of the experiences I have 
thus far had with the instrument, especially since I had an oppor- 
tunity last winter to attach the spectrograph to our new photo- 
graphic telescope also, and thus secure a series of spectrograms 
which permitted a comparison of the results obtained with the 

two refractors. 

366 



THE THIRTY' INCH PULKOWA REFRACTOR 



367 



One of the most important investigations for the objective is 
the determination of its color curve. This had already been done 
by H. Struve, according to Vogel's method, but his observations 
related to only four points in the visual portion of the spectrum, 
and they are also affected by the imperfect achromatism of the 
eye. It was therefore quite interesting to secure a series of spec- 
trograms in order to obtain a more complete idea of the achro- 
matism of the objective. Accordingly spectrograms were made 
on isochromatic and on ordinary plates ; for the former the slit 
was placed in the visual focus of the objective, for the latter in 
the focal planes for Hy and X4410. The measurements were made 
with the Topfer microscope, the spectrogram being placed at 
right angles to the screw, and covered with a solar spectrum 
plate for the purpose of orientation. 

The diameters of the circles of chromatic aberration in the 
visual portion of the spectrum are reduced to the focal plane of 
the rays X6000 — 5000. The following table was obtained by 
platting upon millimeter paper sixteen measured diameters on 
each of the three spectrograms of stars of Classes I and II, and 
then drawing the curve, the hundredth of a millimeter being 
certain : 



X 


Diameter 


X 


Diameter 


X 


Diameter 




mm 




mm 




lU^ 


6000 


0.00 


5400 


0.15 


4800 


0.17 


5900 


0.08 


5300 


O.IO 


4700 


0.37 


5800 


0.15 


5200 


0.05 


4600 


0.61 


5700 


0.22 


5100 


0.02 


4500 


0.88 


5600 


0.23 


5000 


0.0 1 


4400 


1.18 


5500 


0.19 


4900 


0.05 


4300 


1.49 



H. Struve gives the diameters in the planes D-C as follows ; 



X 


Diameter 


6560 
4860 
4340 


0.16 
0.35 
1.77 



It will be seen that these values are greater than those given by 
the spectrograms, but the difference is explained by the fact that 



368 



A. B^LOPOLS/CV 



Struve's measurements are affected by the chromatic aberration 
of the eye. 

The spectrogram on orthochromatic plates (erythrosin) com- 
mences near D, reaches a maximum of intensity at X 5 500, decreases 
to X 5000, and then becomes more intense up to F ; then it dimin- 
ishes as the width increases to /Ty, where the spectrum ceases to 
be measurable. 

The measurement of the diameters on ordinary plates gave 
the following results for the focal plane of X4400 : 



X 


Dianeter 


X 


Diameter 


X 


Diameter 


4270 
4310 
4340 


nm 1 
0.33 
0.20 

0.08 

1 


4380 
4400 
4415 


mm 
0.04 
0.00 
0.03 


4443 
4453 
4668 


0.09 
0.25 
0.56 



The intensity of the spectrum is so slight at X4270 and X4670 
that no lines are visible at those points, and the measurable part 
of the spectrum under ordinary circumstances lies between X4300 
and X4440. 

If the slit is adjusted to the focal plane of X4341, and the 
spectrum of a star of Class I is photographed, the following diam- 
eters are obtained : 



X 


Diameter 


X 


Diameter 


X 


Diameter 


4655 
4550 
448 
4383 


mm 
0.94 

0.73 
0.42 
0.22 


4352 
4341 

4335 
4326 


mm 
0.12 
0.00 
0.05 
0.09 


4308 
4272 
4227 
4217 


mm 
0.16 
0.26 
0.40 
0.48 



It therefore appears that the diameter increases very rapidly 
on each side of Hy in both of these cases, and hence, as well as 
for other reasons to which we shall refer again, the intensity 
decreases rapidly, so that here also the measurable portion of the 
spectrum lies between X4300 and X4450. 

Let us now compare these results with those obtained from 
measurements on spectrograms secured with the photographic 
telescope. On account of the given focal length of the coUima- 



THE THIRTY-INCH PULKOWA REFRACTOR 



369 



tor, only 250"*" instead of the whole 330""" of aperture of the 
objective could be utilized. 



Collimator Settings 47 


Collimator Setting s 45 


X 


Diameter 


X 


Diameter 


4800 
4450 
4340 
4100 
H 


ram 
0.06 
0.0 1 
0.00 
0.02 
0.04 


4800 
4410 
4340 
4200 
4100 
4000 


mm 
0.03 
0.07 
0.06 
0.07 
0.05 
0.00 



We see therefore that in this case the spectrogram is as much 
as three times longer than with the thirty-inch refractor, and this 
depends only upon the fact that the diameters of the circles of 
chromatic aberration are almost constant. It is to be noted 
here that one-half as long an exposure is necessary to secure a 
measurable spectrogram of a star of magnitude 3.5 with the thirty- 
inch as with the photographic telescope. According to the aper- 
tures of the two objectives, however, this ratio should be less 
than one-ninth. 

Let us now examine some of the conditions other than opti- 
cal which affect the performance of the thirty-inch telescope. 
In general the star-images cannot be said to be worse at Pulkowa 
than elsewhere, but the air is for the most part insufficiently 
transparent, — a matter of g^eat importance for the ultra-violet 
end of the spectrum. On this account the spectrum is consider- 
ably weakened even in the region of Hy. That this is not in any 
great degree to be attributed to the glass is seen from the fact 
that there are nights (in Spring) on which the spectrum of a star 
extends to X4270. On other evenings (and such are the major- 
ity) the continuous spectrum of one and the same star is much 
fainter even on the violet edge of Hy than on the red edge, so 
that a systematic error will affect the settings upon this line. In 
most cases the strongest part of the spectrum is in the region 
X 4400-4300, while judging from the diameters of the circles of 
aberration the intensity should be symmetrical on both sides of 



370 A. BtLOPOLSKY 

the smallest circle. For this reason the region mentioned is for 
us the most effective with stars of Class II, the prisms being set 
at minimum deviation for these rays, and the lines of the iron 
spectrum at X4405 and X4415 being used for comparison. 

With the photographic telescope the atmospheric absorption 
is much less felt; the spectra of stars of Class Ila-IIIa suffer 
chiefly, but not nearly as much as with the thirty-inch. 

Let us now more closely examine the mechanical arrange- 
ments of the great refractor, as they are of much importance in 
setting and retaining the star-image upon the slit (o"".03 wide). 
Great inconvenience arises from the unavoidable fact that so long 
and massive a tube cannot perfectly obey the slow-motions in 
right ascension and declination, — and here again the mechan- 
ism of the photographic telescope has a decided advantage. To 
illustrate the difficulty of setting and retaining the star-image 
upon the slit, I may say that the two components of y Virginis 
are very easily confused during the exposure. The observer is 
continually in doubt as to which component is upon the slit, as 
the slightest turning of the declination slow-motion causes the 
image to suddenly spring away from the slit, while the other com- 
ponent takes its place without permitting the eye to actually 
notice the transfer. The star-image is displaced from the slit 
by the very slightest movement, such as a light pressure upon 
the eye-end of the telescope tube (not upon the spectrograph 
itself). It is also very difficult to keep the image within narrow 
limits in the length of the slit (which is parallel to the diurnal 
motion), particularly when there is any wind, and hence the spec- 
trum is always broader than is necessary for the measurements, 
with a corresponding loss in the intensity of the spectrogram. 
When, further, the fact is recalled that in the focal plane for Hy 
the visual star-image is a disk of I'^'^.S diameter, and that the 
center of this disk (the center furnishes the most intense rays for 
the spectrogram) is very hard to locate, as seen in the slit, an idea 
may be gained of the difficulties which arise in making an expo- 
sure for a stellar spectrum, and which often wholly spoil the 
results. 



THE THIRTY' INCH PULKOWA REFRACTOR 371 

It is indeed possible with the thirty-inch to secure spectra of 
stars down to the fourth magnitude by an exposure of one hour 
with a dispersion of two prisms, but this is only true when the 
atmosphere is very transparent, and the star near the zenith. It 
IS also to be remarked that stars with which this is possible must 
belong to Class I, or be pure examples of Class Ila (as a Auriga). 
With stars which are in the transition stage between Classes Ila 
and Ilia (as aTauri, or aBootis), it is impossible to go below 
magnitude 3.5. Even in case of the bright stars, the difference 
in the time of exposure required for white and for yellow stars is 
noticeable : — while the spectrograms of aCygni and a Aurigae are 
sufficiently intense after an exposure of five minutes, those of 
such stars as aTauri require not less than twenty minutes to pro- 
duce similar results. In general we may say that stars down to 
and including magnitude 3.5 are spectrographically accessible to 
the thirty-inch refractor. 

When we think of the results obtained here with the photo- 
graphic telescope, it is perhaps clear enough that the Pulkowa 
thirty-inch, with its present accessories, cannot accomplish one- 
half of what it would be capable with different optical and 
mechanical arrangements, and under a different sky than that of 
Pulkowa, as for instance that of Taschkent or Samarkand, but 
not that of Odessa, the Crimea, or any stations near the sea. 

Pulkowa, Russia, 
February, 1895. 



NOTE ON THE SPECTRUM OF ARGON.« 

By H. F. Nbwall. 

In the course of a spectroscopic investigation in which I have 
been for some time past engaged, a line spectrum, which so far 
as I was able to make out was unknown, has frequently presented 
itself upon my photographs. It appeared in May and June, 
1894, under conditions which led me to call it, for the sake 
of convenience, "the low-pressure spectrum." After their 
announcement at the Oxford meeting of the British Association, 
it seemed for many reasons natural to borrow the first letter of 
Lord Rayleigh*s and Professor Ramsay's names to give to the 
unknown lines and in the measurements of the photographs which 
showed the lines well, there appears an "R" against seventeen 
lines out of sixty-one measured, the remaining lines being known 
to belong to Hg, H, N, and nitrocarbons. It transpires now, as 
I learnt from reading the abstract of the paper in which Lord 
Rayleigh and Professor Ramsay describe their consummate 
researches on argon, that the symbol "A" should have been 
used instead of "R" to designate the lines on my photographs. 
For the lines are Argon lines. 

The conditions under which the spectrum of argon has 
appeared in my investigations are of interest at the present time, 
and I hope a description of them may not be unacceptable. 

A glass bulb was sealed hermetically to a mercury pump of 
the Hagen-Tdpler form, in which there was strong sulphuric acid 
floating on the surface of the mercury. The bulb was exhausted 
as low as'possible and refilled with air. The pressure was reduced 
to about 180 Millionths of an atmosphere (= o°^.i4), at 
which pressure a bright discharge could be passed through the 
residual gases by means of Professor J. J. Thomson's method of 
surrounding the bulb by a coil of wire, which carries a very rap- 
idly alternating current produced by the discharge of a condenser. 

> Read before the Royal Society. 

372 



THE SPECTRUM OF ARGON 373 

The discharge was passed for thirty minutes, during which 
time a photograph of the spectrum was taken. The pressure of 
the gas in the bulb fell during the passage of the discharge from 
the value 174 M (= o"".i3) to 112 M (= o"».o85). The 
spectrum shows the bands of nitrogen strong, also mercury lines 
and nitrocarbon groups strong, hydrogen weak, no oxygen or 
argon. 

Again the discharge was passed for thirty minutes and a 
new photograph was taken. The pressure fell from 100 M 
(=o"".076) to 20 M (=o"^.oi5); the nitrogen spectrum had 
faded considerably, and there had appeared a number of fine lines 
which I was unable, in spite of careful efforts, to identify with the 
lines of any known substances. 

The nature of my method of investigation of spectra is such 
that it is not difficult to pick out of the numerous spectra which 
appear superposed on the photographic plate the lines which 
belong to any one spectrum. 

The results of measurements made in the last few days of 
seventy-two lines in my " low-pressure spectrum " are given 
below, and side by side are given the measurements of the wave- 
lengths determined by Mr. Crookes for the argon lines. 

The agreement of the measurements shows conclusively 
that we have been measuring the same spectrum. Between 
Hy and X3700 the agreement is all that we could hope for, 
taking into account the fact that my measurements were not 
made with a view of giving a final and ca/efully considered set of 
measurements of wave-lengths, but between Hy and Hfi there-is 
a systematic difference of about three tenth-meters, which I am 
unable at present to account for. The agreement of grouping 
and intensity, however, leaves no doubt as to the identity of the 
spectrum of my low-pressure lines with the spectrum of argon. 
I have reduced my measurements with reference to Rowland's 
scale of wave-lengths, and I infer from the value adopted for 
the Hfi (F) line, that Angstrom's scale has been used in Mr. 
Crookes' reduction. The difference between the scales is not 
enough to account for the discrepancies above referred to. 



5^74 



H. F. NEWALL 



H. F. New ALL 




William Crookbs 

January u» liPS 








Thb Two Sfbctva op AacoN 




MBASmtBMBMTS OT LlMBS 








CM PMOrrOGRAFH 


Blub 


RSD 


Wsve-length 


XtmbbAxj 


WBre-length 


Intensity 


Ware.kngdi 


hModtf 


3719.2 


2 


371.80 


4 






3730.0 


8 


372.98 


10 






3738.8 


3 


373.85 


3 






3750.2 


3 










3766.1 


5 


376.60 
377.05 


8 

2 






3781.8 


6 


378.08 
379.95 
380.35 


9 

I 
I 


377.15 




3809.8 


4 


380.95 


4 






3827.0 


— 


382.75 


2 










383.55 


2 


383.55 








384.55 


I 






3850.8 


7 


385.15 


10 






3868.1 


6 


386.85 
387.18 


8 

2 






3873.4 


4 


387.55 


2 






3883.2 


5 


389.20 


5 






3892.2? 




391.50 


I 


390.45 




3918.8 


5 


392.75 


3 






3920.3 


6 










3928.2 


8 


392.85 


9 






3930.8 


3 


393.18 


3 






3932.3 


5 










3944.1 


5 


394.35 


3 










394.85 


9 


394.85 


10 


3968.0 


7 


396.78 


3 






3973.0 


4 










3979.2 


3 


397.85 


I 






3991.3 


4 










3994.8 


6 










4013.8 


8 


401.30 


8 






4033.7 


3 


403.30 


I 






40350 


2 










4038.2 


5 










4042.7 


5 


404.40 


8 


404.40 




4069.7 


2 










4072.4 


9 


407.25 


8 






4075.8 


3 










4082.2 


4 










4104.2 


8 


410.50 


8 






4130.9 


6 


413.15 


3 






4155.8 








415.65 





THE SPECTRUM OF ARGON 



375 



H. F. Newall 




William Crookes 








Juuary 


«4. 1895 






Thb Two Spectra op Argon 




Blue 


Red 


WftTe-length 


iBtensity 


Wavelength 


Intensity 


Wftve-length 


Intensity 




r4i5.95 


10 


415.95 


10 


CN group has, though only 




416.45 


8 


416.45 


4 


o! intensity 5, obliterated 




418.30 


8 


418.30 


8 


this set of lines. 




419.15 


9 


419.15 


9 






419.80 


9 


419.80 


9 








420.10 


10 


420. 10 


10 


4227.5 


8 


422.85 


6 










425.15 


2 


425.15 


3 






425.95 


8 


425.95 


9 


4266.4 


9?N 


426.60 


6 


426.60 


4 






427.20 


7 


427.20 


8 


4277.4 


8?N 


427.70 


3 






4282.1 


6 










4299.4 


4 


429.90 


9 






4308.7 


4 






430.05 


9 


4330.8 


10 


433.35 


9 


433.35 


9 


4336.0 


2 






434.50 


5 


4351.4 


7 


434.85 


10 






4370.4 


8 


436.90 


9 






4375.8 


3 










4379.8 


8 


437.65 


9 






4400.1 


5 










4401.7 


9 


439.95 


10 






4414.1 


4 










4421.2 


4 










4426.0 


10 


442.25 


10 






4431.3 


10 


442.65 


10 






4460.0 


2 










4488.2 


6 


447.83 


6 










450.95 


8 


450.95 
451.40 


9 

2 


4546.5 


7 


454.35 


7 






458X.2 


6 


457.95 
458.69 


6 

6 






4592.0 


8 






459.45 


2 


46x1.0 


9 


460.80 


8 






4632.1 


4 






462.95 


5 


4639.0 


2 










4644.0 


I 










4659.6 


7 


465.65 


5 


470.12 


8 


4729.4 


6 


472.66 


2 






4738.0 


8 


473.45 


6 






4766.6 
4808.0 


5 


476.30 


I 






9 


480.50 


7 






4847.2 


5 


484.75 


I 






4879.8 


5 


487.9 


10 


487.9 


4 



376 H. F. NEJVALL 

The experiments were repeated, with slight variations, several 
times with results which, so far as the spectrum of argon is con- 
cerned, were constant. But it was noted that continued passage 
of the discharge appears to result in the attaining of a certain 
minimum pressure, after which there is slight and slow rise to a 
tolerably fixed pressure. It is not necessary to dwell on these 
points in the present note. 

It is interesting to find argon asserting itself, unsolicited, in 

quite new circumstances, and under conditions which practically 

constitute one more mode of separating argon from nitrogen — 

namely, the getting rid of nitrogen by passing an electric 

discharge through it in the preserfce of hydrogen or moisture 

and acid. 

The Observatory, Cambridge, 
February 14, 1895. 



PRELIMINARY TABLE OF SOLAR SPECTRUM 
WAVE-LENGTHS. V. 

By fisNRY A. Rowland. 







Intensity 






Intensity 


Wave-lciiclh 


SnUtance 


and 
Character 


Wave-length 


SubManoe 


and 
Character 


4414.621 




00 


4420.832 




00 


44M-7I4 


Zr 


00 


4421.090 




000 


4414.896 




000 


4421.290 




OON 


4415.047 


Mn 


2 


4421.495 




00 


4415.199 




foooN 


4421.616 




00 


4415.293 s 


Fc 


1^ 


4421.733 


V 





4415.414 




[oooN 


4421.928 


Ti 


00 


4415.586 




000 


4422.104 




I 


4415.722 




3 


4422.226 







4415.945 




00 N 


4422.461 




00 Nd? 


4416.074 




000 


4422.666 




00 


4416.224 




00 


4422.741 


Fe,Y 


3 


4416.319 




00 


4422.872 




000 


4416.517 




00 N 


4422.985 


Ti 





4416.636 


V 





4423.134 




od? 


4416.81 1 




00 N 


4423.298 


Fe 


I 


4416.985 




2 


4423.430 


Cr 


oN 


4417.163 




000 


4423.630 




000 


4417.275 




ooN 


4423.747 




000 


4417.450 


Ti 





4423.843 




000 


4417.577 


Co 


00 


4424.006 


Fe^ 


2 


4417-740 




00 


4424.233 







4417.884 


Ti- 


3 


4424.368 




00 


4418.044 




000 


4424.457 


Cr 





4418.199 




00 N 


4424.531 




00 


4418.366 




000 


4424.748 




oNd? 


4418.499 


Ti- 


I 


4424.975 




00 Nd? 


4418.590 




00 


4425.321 




000 N 


4418.734 




00 d? 


4425.608 S 


Ca 


4 


4418.944 




ooN 


4425.827 




00 


4419.106 




ooN 


4425.931 




00 N 


4419.265 




ooN 


4426.121 




000 


4419.440 




00 N 


4426.201 


Ti 


oNd? 


4419.675 




00 


4426.536 




000 


4419.768 




000 


4426.617 




000 


4419.944 


Mn 


00 N 


4426.839 




000 Nd? 


4420.100 


V 


ooN 


4427.054 




000 Nd? 


4420.266 




00 N 


4427.266 


Ti 


2 


4420.447 







4427.482 


Fc 


5 


4420.686 


Zr 


00 


4427.623 




0000 



377 



378 



HENRY A. ROWLAND 



Wave-length 



Substance 



4427.760 

4427.875 
4428.083 
4428.256 
4428.438 
4428.711 
4428.873 
4429.077 
4429.366 
4429.456 
4429.664 
4429.804 
4429.958 
4430.070 
4430.221 
4430.356 
4430.524 
4430.646 
4430.785 
4430.929 
4431.302 

4431-453 
4431.525 
4431.660 

4431.785 

4432.009 

4432.247 

4432.330 

4432.477 

4432.587 

4432.736 

4432.904 

4433.089 

4433.208 

4433.390 

4433.554 

4433.742 

4433.948 

4434.057 

4434.168 

4434.361 

4434.504 

4434.605 

4434.810 

4434.918 

4435.1298 

4435.321 

4435.493 

4435.605 

4435.851 s 



La 
La 



V-Cr 



V 
La 
La 
Fe 



Fc 



Cr 
Fc 

Fe 

Fe 
Ti 



Ca 
Fe 



Ca 



Intensity 

and 
Character 



000 
0000 
OON 
OON 

00 N 
Id? 

00 
000 
00 
00 

oooN 
000 N 

00 

00 N 
ooN 
I 

00 
00 

3 



oN 

000 



000 

00 

o 

00 

o 

0000 
000 

I 

ooN 
000 N 
00 N 

3 

00 

ooN 

I 

000 

oNd? 

000 N 

00 



ooN 

000 

5 

a 

ooN 
ooN 
4 



Wave-length 



4436.000 
4436.162 

4436.313 
4436.516 
4436.750 
4436.852 
4436.949 
4437.112 

4437.297 
4437.429 
4437.589 
4437.729 
4437.862 
4438.006 
4438.192 
4438.359 
4438.510 
4438.687 
4438.790 
4438.953 
4439.127 
4439.332 
4439.521 
4439.649 
4439.806 

4439.911 
4440.054 
4440.231 
4440.342 
4440.515 
4440.635 
4440.787 
4440.989 
4441151 
4441.255 
4441.433 
444x591 
4441.718 
4441.881 
4442.127 
4442.241 
4442.421 
4442.510 
4442.579 
4442.751 
4442.842 
4442.996 
4443.161 
4443*365 
4443.459 



Subetantit 



V 

Mn 



Fe-Ni 



Sr. Zr, Ti 
Fc 



Fe 

Fe 
TI 
Zr. 



Fe 
Ti 



Fe 



Fe 
Zr 
Fe 



Intetuity 

And 
Character 



0000 

OOON 



2 

00 

00 

00 

2d? 

000 N 

000 N 

oooN 



00 



00 

00 Nd? 

I 

00 N 

00 

0000 

0000 

oNd? 

00 

0000 



000 

I 

oooN 

00 Nd? 

00 

ooN 

I 





00 

00 

00 

3Nd? 

ooN 

000 

000 

6 

000 

0000 

000 

I 



3 

000 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 



379 



Ware-length 



4443.723 
4443.976 
4444.133 
4444.243 
4444.385 
4444.566 
4444.728 
4444.862 
4445.228 
4445.479 
4445.641 
4445.844 
4446.019 
4446.238 
4446.409 
4446.566 
4446.704 

4446.795 
4447.008 
4447.052 
4447.190 
4447.302 
4447.519 
4447.718 

4447.892 S 

4447.952 
4448.184 
4448.455 
4448.^07 

4449.111 
4449.313 
4449.507 
4449.630 
4449.882 
4450.093 
4450.267 
4450.398 
4450.482 
4450.654 
4450.794 
4450.925 
4451.087 
4451279 
4451.521 
4451.752 
4451.997 
4452.171 
4452.311 
4452.488 
4452.782 



Satetuioe 



Ti 



V-Ti 
Fe, Ti 



Fe 



Fe 

Mn,Fc 

Fe 

Ti 



ZrFe 

Ti? 



Ti 

Mn 
V 



Intensity 

and 
Character 



000 N 

5 

0000 

000 



00 

2 

00 

000 N d? 

000 N 

I 

00 Nd? 

000 N 

0000 N 

00 

00 

0000 

000 

2 

0000 

0000 

2 

00 Nd? 

oooN 

6 

00 

000 N 

000 N 

000 N 

00 N 

2 

00 

00 

00 

00 Nd? 

ooN 

000 

I 

2 

000 

00 

I 

000 N 

000 Nd? 

3 

000 N 

oN 

0000 

000 





Wave-length 



4452.903 
4452.967 
4453.171 
4453.328 
4453.486 
4453.692 
4453.876 
4454.004 
4454.164 
4454.274 
4454.387 
4454.552 
4454.700 
4454.836 

4454.953 S 

4455.193 
4455.342 
4455.485 
4455.615 
4455-710 
4455.815 
4455980 
4456.064 s 

4456.225 
4456.338 
4456.497 
4456.625 
4456.794 s 

4456.945 
4457.040 
4457.207 
4457.330 
4457.435 
4457.600 
4457.712 
4457.835 
4457.940 
4458.110 
4458.239 
4458.409 
4458.550 
4458.690 
4458.850 
4459.003 
4459.199 
4459.301 
4459.525 
4459.670 

4459.779 
4459.922 



Substance 



Mn 

Ti 

Ti 

Fe 



Ca, Zr 

Mn 

Mn. Ti 



Mn 
Ca 



Fe 
Ca 

Mn 



Ti, V, Zr 

Mn 



Fe? 
Mn 

Cr 



Ni 

Fe 

Fe,Cr 



Intensity 

and 
Character 



00 
00 

I 
00 

2 

00 N 

I 

000 

000 

000 

000 

3 

000 

00 

5 

I 

000 
2 

000 N 
000 N 
000 N 
2 

3 

000 
000 
I 

000 N 

2 

000 

000 



000 

00 

2 

2 

0000 

ooN 

000 

2 

2 

000 

o 

000 

00 N 

2 

3 
I 

000 N 
000 
I 



380 



HENRY A. ROWLAND 



Wave-length 



Subtttuios 



4460.063 
4460.164 
4460.276 
4460.389 
4460.462 
4460.525 
4460.700 
4460.944 

4461.093 
4461.242 
4461.365 
4461.545 
4461.592 
4461.818 
4461.983 
4462.165 
4462.365 
4462.525 
4462.621 
4462.750 
4462.860 

4462.933 
4463.060 

4463.152 
4463.300 
4463.425 
4463.569 
4463.698 
4463.843 
4463.997 
4464.138 

4464.389 
4464.503 
4464.617 
4464.844 
4464.938 
4465.069 
4465.143 
4465.295 
4465.385 
4465.519 
4465.667 
4465.775 
4465.975 
4466.147 
4466.333 
4466.415 
4466.548 
4466.727 
4466.886 



V 

Mn 



Mn 

Fc, Zr, Ni 

Fe 

Fc 

Fe-Mn 

V 

Ni 



Ti-Ni 
Ti 



Ti? 
Mn 



Cr 

Ti 



Ni 
Fc 



Intensity 

and 
Characttr 



000 
000 N 
000 
O 

I 

o 

00 Nd? 

00 

000 N 

I 

I 

o 

00 

4 

000 N 

3Nd? 



00 

I 

0000 

00 

00 

000 

00 

00 

00 N 



00 

0000 

0000 N 

0000 

000 

000 

2 

2 

I 

00 

00 

ooN 

oooN 

o 

000 

ooN 

I 

ooN 

ooN 

000 

o 

5 
000 



Wave-length 



4467.0x7 
4467.102 
4467.248 
4467.373 
4467.498 
4467.603 
4467.721 
4467.997 
4468.160 
4468.317 
4468.463 
4468.663 
4468.800 
4468.9X4 
4469x50 
4469.316 
4469.44 X 

4469.545 
4469.73X 
4469.873 

4469.97 X 
4470.x 00 

4470.300 

4470.477 
44^0.648 
4470.799 
4470.875 

447X.OX7 
447X.X66 
4471.250 
447 X. 408 
447X.57X 
447X.724 
447 X. 846 
447X.97X 
4472.076 
4472.241 
4472.37 X 
4472.578 
4472.705 
4472.884 
4472.967 
4473.095 
4473.300 

4473.385 
4473550 
4473.633 
4473798 
4473.927 
4474.001 



Co-Zr 



Cr 



Ti- 



Ti 

Fc 
Co 
V 



Mn 

Ni-Zr 

Ti- 
Ti 



Fe 
Mn 

Ni? 



Character 



000 

I 

000 

000 

00 

0000 

ooN 
00 Nd? 
ooN 
000 N 
ooN 

5 

000 

GOON 

ooN 

I 

0000 

4 

od? 

00 

0000 

000 

X 

000 N 

2 

000 

000 

I 

000 

000 

o 

000 

ooN 

o 

00 

0000 

000 
000 

00 N 

ooN 

I 

o 

o 

000 
000 
000 
000 

000 

00 
00 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 38 1 



Ware-length 



4474.213 
4474.333 
4474.566 

4474.737 
4474.912 
4475.026 

4475.175 
4475.260 

4475.335 
4475.470 
4475.633 
4475.886 
4476.185 
4476.253 
4476.399 
4476.596 
4476.804 
4477.028 
4477.228 
4477.397 
4477.635 
4477.810 

4478.015 
4478.190 
4478.306 
4478.486 
4478.792 
4478.982 
4479.163 
4479.404 
4479.553 
4479.775 
4479.879 
4480.015 

4480.133 
4480.308 
4480.440 
4480.548 
4480.633 
4480.752 
4480.868 
4480.990 
4481.195 
4481.298 
4481.438 
4481.515 
4481.647 
4481.782 
4481.940 
4482.078 



Ti 



Cr 



Fe 
Ag 



Fc-Co 



Mn 
Fe 
Ti 



Fe 
Ti,Ni 

Ti 
Fe 



00 

000 

00 

00 

00 

o 

000 

000 

0000 

00 

oooN 

oooN 

4 

3 

0000 

000 

000 

000 

00 

oooNd? 

ooN 

000 Nd? 

oooN 

o 

000 

ooN 

ooN 

ooN 

ooN 

00 

o 

I 

00 

000 N 

o 

I 

00 

000 

0000 

oN 

0000 

o 

00 

o 

I 

o 

ooooN 

I 

000 N 

oooN 



WsTe*leiigui 



4482.170 
4482.338 
4482.438 
4482.603 
4482.704 
4482.904 
4483.039 
4483.193 
4483.346 
4483.5*3 
4483.706 
4483.825 
4483.942 
4484.078 
4484.250 
4484.392 
4484.555 
4484.667 

4484.859 
4484993 
4485.122 
4485.244 

4485.373 
4485.586 
4485.701 
4485.846 
4486.003 
4486.140 
4486.286 
4486.387 
4486.488 
4486.762 
4486.914 
4487.076 
4487.168 
4487.420 
4487.530 
4487.685 
4487.916 

4488.034 
4488.108 
4488.218 
4488.305 
4488.493 
4488.687 
4488.852 
4488.928 

4489.075 
4489.262 

4489.351 



- Fe 
Fe 



Ti-Fe 
Cr 



Co 
Fe 



Fe 



FeCr 



V 
Fe 
Ti 



ChtractfT 



000 N 
5 

3 

0000 

000 N 

I 

00 

0000 

0000 

0000 

0000 

0000 

o 

o 

000 

4 

000 

00 

000 

000 
000 
000 
000 
000 
000 

3 

000 

o 

000 N 
000 N 
000 N 
00 N 
000 



00 

00 

00 

00 



00 

000 

o 

I 
I 
000 N 

000 
0000 

I. 
o 

2 



382 



HENRY A, ROWLAND 







Intensity 






Intendt, 


WaTc-leneth 


Sobfttaooe 


and 
Character 


Wave-leneth 


Sttbai«>ce 


and 
Character 


4489.505 




00 


4497.268 




000 


4489.630 


Cr 





4497.429 




000 


4489.766 




00 


4497.571 




000 


4489.911 


Fe 


4 


4497 842 


Ti 


ON 


4490.092 




000 


4498.030 


V 


000 


4490.253 


Mn-Fe 


3N 


4498.265 




000 


4490.398 




00 


4498.467 




ooN 


4490.561 




000 


4498.725 




ooN 


4490.701 


Ni 





4498.897 


Cr 





4490.778 




00 


4499.066 s 


Mn 


I 


4490.942 


Fe 


2 


4499.201 




0000 


4490.975 







4499.310 s 




I 


4491.II3 




0000 


4499.525 




oooN 


4491.272 




000 


4499.666 




000 


4491.377 




000 


4499.881 




0000 


4491.570 




2 


4500.122 


Mn 


000 


4491.823 


Cr-Mn 





4500.451 


Cr 





4492.016 


Cr 


00 


4500.537 




00 


4492.139 




000 


4500.669 




000 


4492.278 




000 


4500.807 




00 


4492.475 


Cr. Fe 





4500.932 




000 


4492.700 




00 


450I.II4 




000 


4492.846 




I 


4501.264 


Cr,Mn 





4493.016 




0000 


4501.448 s 


Ti,. 


5 


4493.132 




000 


4501.622 




ooN 


4493.391 




000 


4501.813 




ooN 


4493.543 




00 


4501.946 




oNd? 


4493.695 




T 


4502.157 




ooN 


4493.917 




ooNd? 


4502.217 




0000 


4494.118 







4502.388 


Mn 


2 


4494.222 




I 


4502.603 




00 


4494.356 




00 N 


4502.764 


Fe? 





4494.548 




00 


4502.925 




000 


4494.656 




00 


4503.046 




0000 


4494.738 s 


Fe 


6 


4503^228 




000 


4494.903 




000 


4503.476 




000 


4495.031 




000 


4503.519 




000 


4495.182 


Ti 


00 


4503.654 




000 


4495.426 




00 


4503.926 




00 


4495.590 




ON 


4504.042 


Mn 


00 


4495.738 







4504.224 




0000 


4495.921 




000 N 


4504.371 




oooN 


4496.125 




I 


4504.707 




0000 


4496.318 


Ti 


I 


4504.898 




00 


4496.409 




00 


4505.003 


Fe 


I 


4496.541 




0000 


4505.196 




00 


4496.676 


Mn 


00 N 


4505.404 




ooN 


4496.826 




000 


4505.647 




0000 


4497.023 s 


Cr 


3 


4505.959 




00 


4497.138 


Zr 





4506.092 




000 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 



383 







Inteosit, 






Intensity 


Wcre-Ieasth 


SitlMtaiioe 


and 
Charader 


W«ve.leogth 


Subtttmce 


and 
Chanctcr 


4506.259 


Ba? 


OON 


4514.358 


Fe.Co 


I 


4506.497 




00 


4514.486 




000 


4506.618 




000 


4514.594 




I 


4506.776 




00 


4514.662 


Cr 





4506.907 




00 


4514.817 




000 


4507.009 




00 


4514.957 




ooN 


4507.139 




0000 


4515.134 




000 


4507.266 




0000 


4515.273 




0000 


4507.396 







4515.342 







4507.560 




00 


4515.508 




3 


4507.711 




0000 


4515.606 




00 


4507.919 




00 


4515.763 




ooNd? 


4508.027 




00 


4516.037 




ooN 


4508.177 




00 


4516.255 




000 N 


4508.250 




00 


4516.437 




ON 


4508.455 s 


FcV 


4 


4516.628 




000 


4508.638 




000 


4516.826 




ON 


4508.716 




000 


4517.094 




000 Nd? 


4508.855 







4517.255 




000 


4509.063 




000 


4517.321 


Co 





4509.294 




000 


4517.471 




0000 


4509.456 




ON 


4517.539 




000 


4509.616 







4517.702 


Fe 


3 


4509.758 




000 


4517.764 




000 


4509.904 




I 


4517.923 




000 


4510.I61 




ooN 


4518.004 




000 


4510.344 




0000 


4518.198 


Ti 


3 


4510.434 




0000 


4518.349 




000 


4510.708 




000 


4518.506 




1 


4511.000 







4518.612 




00 


4511.233 




00 


4518.753 







4511*345 


In 


00 


4518.866 


Ti 





4511.516 




ooN 


4519.027 




000 


4511.728 




00 N 


4519.148 




000 


4512063 


Cr 


I 


4519.465 




000 


4512.227 




00 N 


4519.624 




ooooN 


4512.439 







4519.806 




00 


4512.602 




000 


4520.009 




ooN 


4512.663 




000 


4520.157 


Ni 





4512732 




000 


4520.282 




000 


4512.906 


Ti 


3 


4520.397 


Fe?.- 


3 


4513-051 




000 


4520.565 




000 


4513-164 


Ni 





4520.701 




ooN 


4513-385 




000 N 


4520.970 




ooN 


4513-491 




000 


4521.132 




000 N 


4513-603 







4521.304 


Cr 





4513.754 




00 


4521.446 




000 


4513.886 




00 


4521.598 




000 


4514.038 




0000 


4521.834 




oooN 


4514.078 




000 


4522.053 




00 N 



384 



HENRY A. ROWLAND 







Inteniity 






Intendl, 


Wave- length 


Subit»c« 


and 
Chanaer 


Wave* length 


Sabetanoe 


and 
Chancier 


4522.195 




00 


4529.849 


Fe 


I 


4522.286 




00 


4530.020 


Cr 





4522.418 


La 


0000 


4530.172 




oooN 


4522.539 




00 


4530.275 




000 N 


4522.691 


Fe? 





4530.506 




000 N 


4522.802 




3 


4530.668 




00 N 


4522.974 


Ti 


2 


4530.866 


Cr 





4523.H6 




000 


4530.910 


Cr 


I 


4523.250 







4S3M23 


Fe?, Co 


2 


4523.412 




000 N 


4531.327 


Fe 


5 


4523.572 


Mn? 


I 


4531.518 




0000 


4523.751 




000 N 


4531.625 




0000 


4523.910 




000 N 


4531.801 


Fe 


2 


4524.090 




00 


4531.974 




0000 


4524.262 




00 


4532.075 




000 


4524.393 




00 


4532.306 




000 


4524.584 




0000 


4532.485 




000 N 


4524.685 




0000 


4532.743 




000 N 


4524.856 







4532.944 




00 


4525.009 




00 


4533.133 




I 


4525.110 







4533.219 







4525.314 


Fe 


5 


4533.419 


Ti 


4 


4525.412 




000 


4533.583 




0000 


4525.783 




000 N 


4533.710 




000 N 


4526.031 







4533.887 




ooN 


4526.269 


Cr 





4534.139 


Ti-Co 


6 


4526.431 




000 


4534.340 




I 


4526.579 




I 


4534.484 




000 


4526.632 


Cr 


2 


4534.646 




000 


4526.732 


Fe 


1 


4534.788 




000 


4526.887 




0000 


4534.953 


Ti 


4 


4526.95s 




0000 


4535.152 




000 Nd? 


4527.101 


Ca? 


3 „ 


4535.310 


Cr 





4527.332 




000 N 


4535.491 




00 


4527.490 


Ti 


3 


4535.615 




000 


4527.632 







4535.741 


Ti 


3 


4527.807 




000 


4535.879 


Cr 


I 


4527.954 







4535.909 


Zr 





4528.097 




000 N 


4536.094 


Ti 


2 


4528.310 




00 N 


4536.222 


Ti 


2 


4528.473 




000 N 


4536.377 




0000 


4528.647 




ON 


4536.532 




00 


4528.798 


Fe 


8 


4536.675 




00 


4528.929 







4536.853 




0000 


4528.990 







4537.075 




000 


4529.185 




000 


4537.389 




00 N 


4529.394 




000 


4537.592 




00 


4529.482 




000 


4537.845 







4529.656 




I 


4537.986 




000 


4529.728 




I 


4538.138 




ooN 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 



38s 







Intensity 






Intensity 


Wcv«.leiiglli 


Snbttaaoe 


and 
Cbancter 


Wnve-length 


Sabstance 


and 
Chm:ter 


4538.351 




OOON 


4545.861 




000 


4538.539 




00 


4545.983 




000 


4538.634 




000 


4546.129 


Fe.Cr 


3 


4538.770 




ooN 


4546.276 




000 


4538.928 


Fe 





4546.428 




000 N 


4539.013 


Fe 





4546.645 




00 


4539.124 




00 


4546.848 




00 


4539.263 




00 


4546.975 




0000 


4539.424 




00 


4547.101 


Ni 


I 


4539.565 




oooNd? 


4547.192 


Fe 


2 


4539.759 




00 


4547.401 


Ni 





4539.946 


Cr 


oN 


4547.587 




000 


4540.167 




ooN 


4547.675 




000 


4540.385 




00 


4547.815 




000 N 


4540.446 




0000 


4548.024 


Fe 


3 


4540.575 




00 


4548.165 




0000 


4540.672 


Cr 


2 


4548.301 




00 


4540.880 


Cr 


2 


4548.412 




000 


4541.043 




00 


4548.614 




00 


4541.236 


Cr 





4548.756 




00 


4541.352 




000 


4548.938 


Ti 


2 


4541.483 







4549.069 




000 


4541.690 


Cr 


2 


4549.187 




000 


4541.823 




000 


4549.273 




000 


4541.976 




000 


4549.360 




00 


4542.113 




00 


4549 450 




00 


4542.234 




000 


4549.642 


Fe 


2 


4542.400 


Zr 


oN 


4549.808 


Ti-Co 


6d? 


4542.600 


Fe 


iN 


4549.990 







4542.785 


Cr 





4550.162 




oooN 


4542.876 







4550.293 







4543.012 




0000 


4550.445 




00 


4543.202 




000 


4550.601 




000 


4543.402 




00 


4550.743 




000 


4543.525 




0000 


4550.942 


Fe? 


2 


4543.900 




00 


4551.139 




000 


4543.990 


Co 





4551.261 




000 


4544.190 




I 


4551.399 


Ni 





4544.365 




000 


4551.458 




000 


4544.444 




0000 


4551.691 




000 N 


4544.655 




ooN 


4551.824 







4544.788 


Cr 


1 


4552.018 




00 Nd? 


4544.864 


Ti 


3 


4552.314 







4545.008 




00 


4552.460 







4545.142 




ooN 


4552.632 


Ti 


2 


4545.311 




I 


4552.725 


Fe 


I 


4545507 


Cr-V 





4552.824 




000 


4545.568 




00 


4553.065 




000 N 


4545.713 






4553.219 




00 


4545.770 




000 


4553.346 


Ni 






386 



HENRY A. ROWLAND 







Intensfty 






Intensity 


Wave-length 


Substance 


and 
Character 


Wavelenetb 


Siibttance 


and 

Character 


4553.546 




000 Nd? 


4562.541 







4553.796 




0000 N 


4562.649 




0000 


4554.009 




0000 


4562.814 


Ti 


00 


4554.211 «s 


Ba 


8 


4563.059 




000 N 


4554.423 




0000 


4563.4 U 




00 


4554.484 




0000 


4563.599 


Ti 


00 


4554.626 




I 


4563.939 S 


Ti 


4 


4554.707 




000 


4564.069 




00 


4554.869 




000 N 


4564.203 




0000 


4555.007 




00 


4564.352 




00 d 


4555.162 




2 


4564.5" 




00 


4555.264 




00 


4564.629 




0000 


4555.468 




ooN 


4564.750 




00 


4555.662 


Ti 


3 


4564.875 







4555.830 




000 


4565.002 


Fe 





4555.910 




00 


4565.215 




oooN 


4556.063 




3 


4565.296 




ooN 


4556.306 


Fe-Cr 


4 


4565.488 


Fe 





4556.549 




0000 N 


4565.597 




00 


4556.719 




ooooN 


4565.688 


Cr 


3 


4556.934 




0000 N 


4565.842 


Co-Fe 


2 


4557.107 







4565.905 




00 


4557.262 




000 


4566.031 




000 


4557.457 




oN 


4566.198 




ooN 


4557.689 




0000 N 


4566.414 




ooN 


4557.927 




00 


4566.555 




0000 


4558.060 




00 


4566.693 


Fe 


I 


4558.162 




0000 


4566.834 






4558.285 







4567.046 


Fe 


I 


4558.402 




00 


4567.168 




00 


4558.640 




00 


4567.345 




0000 


4558.827 


Cr? 


3 


4567.391 




0000 


4558.949 




000 


4567.584 




ooN 


4559.106 




0000 


4567.755 




000 


4559.526 




0000 


4567.917 




0000 


4559.728 




0000 


4568.221 






4559.980 




0000 


4568.499 







4560.102 


Ni, Ti 





4568.777 




00 


4560.266 


Fe 


2 


4568.940 


Fe 


I 


4560.457 




00 


4569.034 







4560.589 




00 


4569.243 




ooN 


4560.740 




000 


4569.425 




000 N 


4560.892 




00 


4569.536 




ooN 


4561.044 




00 


4569.693 


Cr 


00 


4561.145 




00 


4569.788 


Cr 





4561.368 




ooN 


4569.992 






4561.591 




I 


4570.199 




ooN 


4561.909 




ooN 


4570.559 




oooN 


4562.148 




ooooN 


4570.781 




000 Kd? 


4562.406 




0000 


4571.095 




ooN 



' This line is either double or reversed in the Sun. The Ba line seems to coincide 
with the center. 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 



387 







Intensity 






Intensity 


Wave-length 


Sobstaaoe 


•nd 
Character 


Wave-leniKth 


Sttbatanoe 


and 
Character 


4571.275* 


Mg 


5 


4580.080 




000 


4571.470 




oooNd? 


4580.228 


Cr 


3 


4571.618 







4580.326 




0000 


4571.720 




0000 


4580.463 




000 


4571.849 


Cr 


I 


4580.590 


V 


I 


4571.976 




00 


4580.762 


Fe-Ni 


I 


4572.156 s 


Ti. 


6 


4580.911 




000 


4572.366 




000 


4581.055 




ooN 


4572.457 




00 


4581.221 




ooN 


4572.600 




0000 


4581.369 







4572.766 




00 


4581.575 


Ca 


4 


4573.041 




oN 


4581.693 


Co.Fe 


4 


4573.231 




0000 


4581.805 




00 N 


4573.828 




0000 


4582.007 




000 Nd? 


4573.960 


Ba? 


000 


4582.250 




000 Nd? 


4574.168 




00 


4582.483 




oN 


4574.396 


Fe 


I 


4582.683 




ooN 


4574.537 




0000 


4582.851 




000 N 


4574.658 




ooN 


4583.011 




I 


4574.739 




0000 


4583.126 




00 


4574.899 


Fc 


2 


4583.296 




00 


4575.075 




ooNd? 


4583.424 




000 


4575.286 




00 


4583.587 







4575.402 




0000 


4583.749 




00 N 


4575.600 




00 N 


4583.892 




00 


4575.726 


Zr 


00 N 


4584.018 


Fc- 


4 


4575.964 







4584.168 




00 


4576.096 




0000 


4584.269 




00 


4576.268 




000 Nd? 


4584.447 




00 N 


4576.512 




2 


4584.618 




000 


4576.686 




00 


4584.730 




0000 


4576.769 




OQO 


4584.900 


Fe 


1 


4576.957 




000 


4585.001 




2 


4577.X81 




000 Nd? 


4585.x 18 




00 


4577.356 


V 





4585.257 




00 


4577.501 




000 


4585.368 




000 


4577.656 




000 N 


4585.519 







4577.868 




00 


4585.772 




000 


4577.988 




000 


4585.874 




00 


4578.220 




00 Nd? 


4586.047 


Ca 


4 


4578.503 




00 


4586.155 







4578.732 s 


Ca 


3 


4586.315 




00 


4578.909 


V 


ooN 


4586.408 


Cr 


I 


4579.062 




ooN 


4586.552 


V 


I 


4579.231 




ooN 


4586.716 




000 


4579.359 




000 


4586.896 




000 N 


4579.506 







4587.X69 




00 


4579.684 




00 Nd? 


4587.308 


Fe 


2 


4579.862 


Ba? 


00 


4587.571 




00 Nd? 


4579.994 







4587.777 




00 N 



388 



HENRY A. ROWLAND 



Wave-lengtb 



Substance 



4587.898 

4588.052 

4588.180 

4588.381 s 

4588.574 

4588.697 

4588.859 

4589.193 

4589.468 

4589.686 

4589.912 

4590.126 s 

4590.246 
4590.388 
4590.516 
4590.664 
4590.851 
4590.965 
4591. 1 17 
4591.290 
4591.421 
4591.574 
4591.693 
4591.91 1 
4592.027 
4592.231 
4592.393 
4592.534 
4592.707 
4592.840 
4592.990 
4593.102 
4593.355 
4593.543 
4593.704 
4593.883 
4594.002 
4594.113 
4594.297 

4594459 
4594.590 
4594.820 
4594.964 
4595.067 
4595.227 
4595.386 
4595.540 
4595.651 
4595.770 
4595.865 



Cr 



Cr- 



Ni 
Fe 



Co 

Ni 

Fe 
Cr 



Intensity 






Intensity 


and 


Wave-length 


Snbstance 


and 


Character 






Chaiacter 


00 


4596.128 


Ni 





OOON 


4596.245 


Fe 


2 


OON 


4596.414 




000 


3 


4596.589 


Cr- 


I 


ooN 


4596.753 




00 


00 N 


4596.857 




000 


00 


4597.080 


Co 


ON 


00 N 


4597.211 




00 


ooN 


4597.430 







000 


4597.560 







000 Nd? 


4597.775 




000 N 


3 


4597.929 




I 


0000 


4598.050 




I 


000 


4598.186 




0000 


000 


4598.303 


Fe 


3 


oooN 


4598.456 




0000 


000 


4598.547 




00 





4598.612 




00 


00 


4598.792 




000 


00 


4598.918 







00 


4599.109 




000 


2 


4599.183 




00 


I 


4599.408 


Ti 


00 


000 N 


4599.618 




0000 


ooN 


4599.752 




000 


I 


4600.018 


Fe? 


2 


00 


4600.145 




000 


00 


4600.279 


Cr 


X 


2 


4600.383 




ooN 


4 


4600.541 


Ni 


2 


0000 


4600.737 




ooNd? 


000 


4600.932 


Cr 


3 


00 


4601. XI4 







000 


460 X. 207 


Cr 





I 


4601.319 




00 


000 


4601.452 




00 





4601.557 




00 





4601.734 




000 


2N 


4601.917 




000 


000 N 


4602.0x3 




000 


ooNd? 


4602.183 s 


Fe 


3 


ooN 


4602.356 




0000 


0000 


4602.565 




0000 





4602.717 




0000 


00 


4602.929 




0000 


00 


4603.126 


Fe 


6 


2 


4603.282 




000 


00 


4603.525 










4603.665 




0000 


000 


4603.799 




00 



TABLE OF SOLAR SPECTRUM WAVE-LENGTHS 



389 



Ware-leacth 



4603.902 
4604.031 
4604.137 
4604.420 
4604.580 

4604.735 
4604.863 
4605.025 

4605.171 
4605.278 

4605.430 
4605.536 
4605.640 
4605.769 
4606.017 
4606.189 
4606.404 
4606.574 
4606.687 
4606.969 
4607.274 
4607.394 

4607.510 8 

4607.687 

4607.831 

4608.037 

4608.305 

4608.406 

4608700 

4608.887 

4609.023 

4609.447 
4609.540 

4609.752 
4609.833 
4610.088 
4610.267 
4610.365 
4610.771 
46II.II8 
4611.249 
4611.368) 

4611.469)* 

4611.664 

46II.816 

4612.000 

4612.138 

4612.251 

4612.446 

4612.646 



Fc? 

Ni 

Mn 

Ni,C 



Sr 
Fe 



Cr 
Fe 



Cr 



Itttcnsity 

and 
Chancter 



00 

O 

O 

OON 

0000 

2 

000 

00 

3 

000 

000 N 



00 

2 

ooN 

000 N 

2 

ooN 

000 N 

oooN 

ooN 

0000 

I 

000 N 

4 
000 

000 

0000 

000 Nd? 

ooN 

0000 N 

o 

000 

ooooN 

oooN 

o 

0000 

o 

oooN 

000 

00 

o 

5 

000 

00 

000 

00 

000 

000 

000 



Wave-lengtb 



4612.794 
4612.925 

4613137 
4613.386 

4613.544 
4613.738 
4613.892 

4614.097 
4614.38S 
4614.529 

4614.713 
4614.761 
4614.914 
4615.II4 
4615.427 
4615.632 

4615.743 
4615.896 
4616.II4 
4616.305 
4616.472 
4616.644 
4616.804 
4616.923 

4617.134 
4617.244 
4617.452 
4617.636 
4618.046 
4618.150 
4618.303 
4618.536 
4618.688 
4618.971 

4619.133 
4619.285 
4619.468 
4619.607 
4619.71 1 
4619.852 
4619.963 
4620.072 
4620.313 
4620.522 
4620.693 
4620.986 
4621.208 
4621.299 
4621.482 
4621.654 



Fe 
Cr, La 



Fe? 



Fe? 
Cr 

Ti 



Fe- 

Fc 
Cr 



>e 



Inteasity 

and 
Character 



00 
000 
00 N 

3 

3 

000 Nd? 

00 

I 

1 

0000 

00 

00 

00 

0000 N 

0000 N 
000 

I 

00 

00 

4 

000 
ooN 

1 N 
000 
000 
000 N 

3 

000 N 

00 

00 

0000 

00 

00 

4d? 

000 N 

ooN 

3 

0000 

I 

0000 

00 

0000 

00 

ooN 

I 

000 N 

0000 

000 

0000 N 

00 



390 



HENRY A. ROWLAND 







Inieasity 






Intensity 


Wave-length 


StttMUBce 


and 
Character 


Wavelength 


Siibatanoe 


and 
Character 


4621.795 




00 


4630.306 


Fc 


4 


4621.947 




•000 


4630.582 




oooN 


4622.065 


Cr 





4630.740 







4622.128 


Cr 


I 


4630.958 




ooN 


4622.307 




000 N 


4631.212 




oN 


4622.433 




OOON 


4631.385 




000 


4622.627 


Cr 


I 


4631.512 




0000 


4622.733 




000 


4631.663 







4622.929 


Cr 





4631.900 




ooNd? 


4623.071 




000 


4632.129 




00 Nd? 


4623.279 


Ti 


2 


4632.320 


Cr 





4623.476 




ooN 


4632.503 




ooN 


4623.759 







4632.654 




ooN 


4624.053 




oooN 


4632.825 




ooN 


4624.265 




00 


4632.991 




1 


4624.443 




000 N 


4633.100 


Fe 


4 


4624.594 




00 N 


4633.272 




ooNd? 


4624.741 




00 N 


4633.432 


Cr 





4624.923 




00 N 


4633.555 




oooN 


4625.074 




00 N 


4633.722 




ooN 


4625.227 


Fe 


5 


4633.950 




oN 


4625.378 




000 N 


4634.187 




0000 


4625.489 




ooN 


4634.254 




2 


4625.613 




ooN 


4634.441 




000 


4625.947 




ooN 


4634.547 




000 


4626.096 


Cr 


ON 


4634.780 




000 


4626.198 




000 


4634895 


Fc? 


I 


4626.358 


Cr 


5 


4635.048 




00 


4626.532 




ooN 


4635.210 




000 


4626.718 


Mn 





4635.352 


V 


ooN 


4626.825 




000 


4635.489 







4626.975 




00 


4635.598 




0000 


4627.190 




000 N 


4635.736 




00 


4627.392 




00 N 


4635.803 







4627.547 







4635.884 




0000 


4627.726 







4636.027 


Fe 


2 


4627.829 




0000 


4636.192 




oooN 


4628.032 




0000 


4636.339 




ooN 


4628.197 




0000 


4636.501 







4628.335 







4636.740 




0000 


4628.448 






4636.851 




00 


4628.637 


Cr 


00 


4637.108 




000 


4628.860 




00 


4637.221 




00 


4629.092 




ooN 


4637.352 


Cr 





4629.248 




00 


4637.474 




00 


4629.521 8 


Ti-Co 


6 


4637.^5 « 


Fe 


5 


4629.714 




ooN 


4637.845 




0000 


4629.844 




000 


4637.938 


Cr 





4629.979 




00 


4638.050 


Ti 


00 


4630.125 




000 Nd? 


4638.193 ? 


Fe 


4 



TABLE OF SOLAR SPECTRUM WA VE-LENGTHS 



391 



^ftTC*iCllftll 



4638.709 
4638.879 
4639.132 
4639.353 
4639.538 
4639.679 
4639.846 
4640.119 
4640.280 
4640.468 
4640.681 
4640.883 

4641.147 
4641.390 
4641.693 
4641.851 

4642.179 
4642.306 
4642.424 
4642.765 
4643.005 
4643.235 
4643.389 
4643.475 

4643.645 s 

4643.912 

4644.066 

4644.572 
4644.703 
4645.368 
4645.483 
4645.671 
4645.819 
4645.965 
4646.058 
4646.168 
4646.347 
4646.552 
4646.676 
4646.815 
4646.962 
4647.145 
4647.354 
4647.454 
4647.617 
4647.876 
4648.135 

4648.297 
4648.497 
4648.591 



Snbstanoe 



Ti 
Cr 
Ti 
Ti 



Fc 



Ti 



Cr 
Cr 
Cr 

Fe 
Cr 



Intensity 

and 
Character 



00 

00 N 

o 
ooN 

2 
o 

2 
I 
000 

I 

00 
0000 

ooNd? 

o 

oooNd? 

000 

00 

00 

00 

ooN 

00 

000 N 

00 

00 

4 

ooN 

oooN 

ooNd? 

00 

o 

000 

oN 

oooN 

00 

000 

000 

5 

o 

00 

I 



oooN 

00 

o 

4 

oooN 

I 



000 

000 



Ware-length 



4648.835 S 

4649.031 

4649.124 

4649.337 
4649.476 
4649.613 
4649.819 
4649.992 

4650.193 
4650.296 
4650.488 
4650.725 
4650.985 
4651. 121 
4651.290 
4651.461 
4651.685 
4652.045 
4652.198 

4652.343 
4652.447 
4653069 
4653.216 
4653.323 
4653.485 
4653.551 
4653.677 
4653.819 
4653.960 

4654.077 
4654.218 

4654.327 
4654.478 
4654.672 
4654.800 
4654.912 
4655.419 
4655.634 
4655.832 
4655.967 
4656.127 
4656.228 
4656.365 
4656.481 
4656.644 
4656.815 
4656.992 
4657.154 
4657.380 
4657.554 



Substance 



Ni 
Cr 



Cr 
Ti 

Cr 
Cr 



Fe 
Fe 
Cr 



Ti 

Ti 
Cr 

Ti 



Ti? 



Intensity 

and 
Chanaer 



4 

00 

00 

ooN 
000 
o 
oNd? 

o 
o 

000 

o 

00 
00 
00 
00 

4 

0000 N 
0000 N 

0000 

5 

0000 

0000 

0000 

0000 

000 

00 

00 

000 

000 

000 

000 

o 

ooN 

4 

5 

00 

ooN 

000 Nd? 



o 

0000 

o 

o 

000 

3 

000 
000 
I 

2 
00 



392 



HENRY A. ROWLAND 



Wave-lengtb 



Sttbttanoe 



4657.625 
4657.766 
4658.034 
4658.217 
4658.343 
4658.475 
4658.675 
4658.827 
4659.054 
4659.338 
4659.547 
4659.707 
4659.940 
4660.145 
4660.247 
4660.412 
4660.602 
4660.801 
4660.902 
4661.083 
4661.327 
4661.507 
4661.712 
4661.965 
4662.149 
4662.278 
4662.390 
4662.496 

4662.693 
4662.930 

4663.354 
4663.492 
4663.587 
4663.734 
4663.882 

4663.999 
4664.139 
4664.358 
4664.497 
4664.720 
4664.965 
4665.355 
4665.430 
4665.722 
4665.852 
4665.998 
4666.076 
4666.279 
4666.387 
4666.526 



Fe? 



Cr 
Co 



Cr 



Cr 



Cr 
Cr 



Intensity 

and 
Character 



000 

I 

OON 

000 

0000 



ooN 

00 N 

00 Nd? 

000 Nd? 

000 N 

000 N 

ooooN 

ooN 

ooN 

00 



000 

00 

o 

00 N 

ooN 

I 

ooN 

I 

000 

0000 

0000 

o 

oN 

o 

I 


000 

o 
I 



ooN 
ooN 
00 Nd? 

3 

00 

000 

00 

000 

000 

I 





000 



Wav6>Ien£tli 



666.655 

666.782 

666.925 

667.066 

667.159 

667.339 

667.424 

667.626 

667.768 

667.941 

668.099 

668.243 \ , 

668.331 ) • 

668.550 

668.749 

669.017 

669.164 

669.354 

669.504 

669.568 

669.700 

669.829 

670.001 

670.155 
670.346 
670.590 
670.732 
670.915 
671.080 
671.222 
671.388 
671.601 
671.742 
671.858 
672.087 
672.209 
672.370 
672.509 
672.710 
672.805 
673.012 
673.144 
673.347 
673.460 
673.617 
673.818 
673.962 
674.131 
674.275 
674.484 



Cr 



Ni 
Cr 

Fe 

Ti 
Ni 



Fc 



Fc 
Cr 



Ni 



Mn 



Fe 



Character 



I 
O 

I 

ooN 
I 

00 
o 

4 
3 
I 
000 

2 

4 

ooN 
iN 
oooN 
000 N 

3 

I 

000 

000 

00 

000 N 

000 N 

I 

2 

00 * 

000 

00 

00 

000 

I 

00 

o 

000 

00 

000 

3N 

000 

000 

I 

000 

4 

I 

ooN 

ooN 

ooN 

0000 

iN 

o 



ON MARTIAN LONGITUDES. 
By Percival Lowell. 

With the object of constructing a map of the planet, I made, 
during last October and November, the following series of obser- 
vations on the longitudes and latitudes of prominent points on 
the disk of Mars. The observations cover thirty-six points in 
all, and were taken between October 12 and November 22. 

The longitudes were measured with a power of 440 on the 
micrometer of the 18-inch glass. The longitudinal thread of the 
micrometer I adjusted parallel to the polar axis of the planet 
and then noted the instant of the meridian passage of the point 
to be determined. 

Between the eye and the eyepiece I placed yellow glass. 
The interposition of a piece of suitably tinted glass — deep 
yellow seems the best, as theoretically it should be, owing to 
the correction of the glass for light of that color — reduces the 
size of the spurious disk made by each point of the planetary 
one, and so steadies the image and draws out the detail. I 
discovered subsequently that Schiaparelli had made use of this 
same device. 

The latitudes I got by estimating the positions of the 
points upon the polar axis, at or near the times of meridian 
passage. 

The latitudes are necessarily not so trustworthy as the longi- 
tudes. Of the accuracy of the best of the latter the probable errors 
are witness. The values given are those after all corrections due to 
the phase have been taken into account. Such correction is not so 
simple as it might be, owing to the fact that the phase axis and 
the polar axis did not in general coincide. The amount of the 
lacking lune had therefore to be calculated, both for differing 
points on the disk and for different days of observation. The 
values are given to tenths of a degree, greater accuracy being 

illusory. 

393 



394 



PERCIVAL LOWELL 





Fastigium Aryn. 








LONG. 


WT 


Oct. 12 


- 


- - 4^9 


I 


15 


- 


• - 4^8 


I 


16 


- 


2^2 


I 


17 


- 


- o°.7 


I 


Mean 


- l\2. 




Probable 


error ±©".7 






Fastigium Aryn. 




Nov. 17 


- 


- ■ 5M 


I 


18 - 


- 


- 4^7 


I 


19 


- 


- 4".9 


I 


20 - 


- 


- 4".7 


I 


21 


- 


- - 3°.9 


% 


22 - 


- 


- 3^4 


% 


Mean 


- 4".7 




Probable 


error ±o'*.2 





Nov. 6 
9 



10 
II - 
12 
13 - 



Mean of both sets 
weighted ac- 
cording to their 
probable errors 4*. 6 
Probable error ±o\2 

SoLis Lacus (center). 

- 90^7 
92'.4 

- 90''.4 

92".2 

- 90^.3 

91 '.8 



Mean - - 91 '.3 
Probable error dbo'.3 

Phcenix Lake. 
Nov. 5 - • - . ,i,<»2 
6 - - - lU^'.S 
9 - - -III*'. 

10 - - II2^6 

Mean - - ii2'.8 
Probable error rfco'.s 



z 

I 
3 
3 
3 
3 



Beak of the Sirens. 

LONG. 

Nov. 4 - . - ,26-.2 

9 - - - I27'.7 

10 - - . i25'.8 



Mean 
Probable error 



I26'.6 

±o\4 



Sinus Titanum. 



Oct. 31 
Nov. 2 

3 
4 
5 



Mean - 
Probable error 



174'. 
■ I74'.4 

I73^7 
• 174^6 

I75".6 

I74'.5 

±0\2 



Oct. 20 
21 



Syrtis Major. 

291'. I 
- 293«. 

292 •.4 
dz0'.6 



Mean - 
Probable error 



Oct. 19 - 
21 - 
Nov. 20 - 



Hammonis Cornu. 

- 32o^I 
- 3I9'.6 

- 319*.! 



Mean - - 319^.6 
Probable error ±o**.2 



Nov. 17 
18 

19 
20 
21 
22 



Edom Promontory. 

- 357^6 
- - 358".8 

- 357'. 

- ■ 358'.5 
- - 356^5 

- - 355^9 



Mean - - 357«.6 
Probable error ±o\3 



WT. 

I 
2 
2 



3 
5 

2 

3 
4 

2 



PLATE XVIII 




s 

MARS 

F'astigium Aryn at October Presentation 



MARTIAN LONGITUDES 39 S 

The first fact that emerged from these observations was that 
all the longitudes as given in Marth's ephemeris were affected by 
a systematic error of about 5^. In other words the meridian 
passage with regard to the Earth of any Martian meridian 
occurred invariably some twenty minutes later than the time set 
for it to do so by Marth. In Mr. Marth's times all corrections, 
such as the equation of light, had already been allowed for, and 
in the observed passages all corrections for phase were, of course, 
similarly considered. All the data were, therefore, presumably 
correct, and the discrepancy was unmistakable. All the Martian 
longitudes were the very palpable amount of twenty minutes 
behind time. 

As early as June, when the first drawings were made here at 
this opposition, it was evident that something was wrong with 
the longitudes, as those deduced from Marth's ephemeris for 
the time at which any given drawing was made did not coincide 
with those taken from Schiaparelli's chart. 

So soon as I began special observations upon the longitudes 
the cause of the discrepancies became clear. The Fastigium 
Aryn at its October presentation was the first point I timed, and 
I found that in spite of errors of observation it never once suc- 
ceeded in passing the center of the disk as early as the time pre- 
dicted for it in the ephemeris. 

When the Sinus Titanum came round in November, I 
found it similarly to be behind time. So with other prominent 
points. Measures of the Fastigium Aryn at the November pres- 
entation confirmed the previous tardiness, and showed the 
amount of the error to be 4^.6 of longitude. This is a discrep- 
ancy far transcending the probable error of the observations as 
deduced from their discussion. 

As all the observations are independent, their mutually con- 
firmatory character is conclusive as to their individual trust- 
worthiness, and to their disclosure of some systematic error in 
the ephemeris. 

Reduction of the observations gives for the longitude 
of the Fastigium Aryn 4^.6; that being the mean of both 



396 PERCIVAL LOWELL 

the October and the November determinations, when the mean 
of each set is weighted according to its own probable error. 
The probable error of the result is o**.2. 

In order to compare this value of the longitude of the zero 
meridian with that got by Schiaparelli in 1879, we must take 
into account the latitudes of the point observed on the two occa- 
sions. For the Fastigium Aryn is the tip of a tapering isosceles 
triangle that projects into the Sabaeus Sinus, and the axis of the 
triangle does not lie due north and south, but is inclined 
about 15** to the meridian in the direction N. N. E. and S. S. W. 
The shape of the promontory is thus an entering wedge to some 
uncertainty. For, owing to the inclination, the longitude of the 
point measured will depend upon the latitude taken for it. 
Indeed, at certain seasons, the position of the extremity of the 
peninsula is further masked by a sort of causeway that makes 
out southward till it eventually jolYis Deucalionis Regio, thus 
cutting the Sabaeus Sinus completely in two. The continuation 
first became evident this year in October. To be sure, therefore, 
that the same point is measured on different occasions, the lati- 
tudes must agree ; otherwise some divergence in longitude will 
result simply from such latitude in the observations. Now, in 
1879, the point on the peninsula measured by Schiaparelli lay in 
N. 1° ; this year what was taken as the tip turned out to lie in 
N. 2*^.8. Consequently the two points determined differed in 
longitude by the difference of latitude, 1^.8, into the sine of 15^, 
or by 0*^.47. Adding this amount to the above 4°.6 we get 5°.i 
as the longitude this year of the point of the Fastigium Aryn 
found by Schiaparelli in 1879 to be in longitude 0^.92. These 
two amounts are therefore respectively to be subtracted from 
all the other longitudes in their sets in order to compare the two 
sets together. 

When this is done the following close agreement appears 
between certain points determined by me this year and the cor- 
responding ones determined in 1879 by Schiaparelli. 



PLATE XIX 




MARS 
Sinus Titanum at November Presentation 



MARTIAN LONGITUDES 



397 



FMticiaaAiyB 


Gaagct 




SiBQsTiuunim 


1879 ©• 

1894 0* - - - - 


55^3 
54^6 


107-. 
107'.7 


169 %a 
I69^4 



CiBunerimD Mare (Mouth of the iethiops) 



1879 239». 
1«94 a37'.9 



(Mouth ol) 



Enphnm 
(Mouth oO 



335".4 
334".3 



337' 
338' 



It may also be interesting to note how like are the values of 
the latitude of the Sinus Titanum taken at the same epochs. 

Latitude Sinus Titanum 1877 S. 18''. 17 

" " 1879 S. I9'.33 

" " 1894 S. 20*.0 

The cause that at once suggests itself for such discrepancy 
between the calculated and the present observed positions is that 
the received time of rotation of the planet is a trifle too small, 
and that the longitudes in consequence are falling slowly behind 
their predicted times of meridian passage. That there is any 
error in the computation of the ephemeris is, with so admirable 
a computer as Marth, practically out of the question. Further- 
more a preliminary search just made by him in consequence of 
these observations of mine reveals none. 

It is interesting to note that this increase in the Martian 
longitudes, like many other astronomical matters, has been 
observed without being recognized before, and by more than one 
observer. 

At the opposition of 1892 Keeler ("Physical Observations 
of Mars made at the Allegheny Observatory in 1892," in the 
Memoirs of the Royal Astronomical Society) found, on comparing 
his drawings meridianed by Marth ephemeris with photographs 
of a globe made by him from Schiaparelli's chart and set to the 
longitude and latitude of the time of observation : 

"that there was a very satisfactory agreement among themselves in the 
positions of the markings on the various drawings, but that there was a small 
and nearly constant difference of longitude between the drawings and the 
photographs of the globe, the longitude of the central meridian on the latter 



398 PERCIVAL LOWELL 

exceeding that of the drawings by about seven degrees. I therefore made 
another set of photographs, with slightly different positions of the globe, 
according to the following plan : The axis of the globe was set to the proper 
inclination so that the latitude of the center of the image was that of the center 
of Mars at the time of observation. The globe was turned on its axis until 
the image on the ground glass was in the best general agreement with the 
drawing made at that time, and the image was then photographed as before. 
The agreement of the drawings and photographs in their main features was 
then remarkably close. 

" I am unable to account for this constant difference of longitude. The 
most natural explanation is that it is due to constant error in estimating 
the position of the diameter of a large disk, but, according to other experi- 
ments, my personal error of estimated bisection is too small to account for the 
difference.*' 

A similar systematic increase in all the longitudes was found 
in 1890 by Wislicenus at Strasburg, of which in La Planite Mars 
Flammarion simply says : "Ces positions nous paraissent toutes 
un peu trop a droite." 

Both sets of longitudes were a little too much to the right, 
not because either set of observations was wrong, but because, as 
I think is now evident, the received time of rotation is too small. 

A second point to which I wish to draw attention is the rela- 
tive desirability of different Martian markings as zero meridians. 
Of these there are three that commend themselves specially for 
such purpose, the Fastigium Aryn, the Lacus Phcenicis and the 
Sinus Titanum. Each of the three possesses certain advantages 
over the other two, and certain corresponding disadvantages. 
In the first place all are affected in visibility and, therefore, in 
effectiveness for the present purpose by the seasonal change 
that sweeps across the face of the planet. Before the summer 
solstice of the southern hemisphere, the Sinus Titanum is much 
more difficult to identify than it afterwards becomes. This is 
owing to the obliteration early in the season of Atlantis. Before 
this peninsula has differentiated itself from the Mare Sirenum 
on the one side and the Mare Cimmerium on the other, the 
Sinus Titanum is itself inconspicuous, being simply a depression 
in a more or less straight coast line. After Atlantis has appeared, 
however, the Sinus Titanum is perhaps the most prominent mark- 



PLATE XX 




N 

MARS 
Lacus Phcenicis at November Presentation 



MARTIAN LONGITUDES 399 

ing on the disk. For a zero meridian, therefore, it is not good 
before the summer solstice of its hemisphere and excellent after- 
ward. This was certainly the case this year, and if we may judge 
generally by this year's changes, which I think we may, should 
be true at subsequent oppositions. 

On the other hand the Lacus Phcenicis proved this year better 
early in the season than late. At the end of August, that is 
about the time of the summer solstice of the Southern hemi- 
sphere, it was more conspicuous than it turned out to be in Novem- 
ber. As the observations taken on it for longitude were made 
at the latter date, its position is not so satisfactory as that of the 
Sinus Titanum. Its probable error comes out some five times as 
great as that of the latter marking. 

With the Fastigium Aryn lateness in the season also proved 
a drawback. Although I got a better result in November than 
in October, this was due not to the intrinsic ease of the later 
observations, but to the greater care with which the latter were 
taken. The fading out of the region of the Sabaeus Sinus between 
its October and November presentations was most marked, and 
very unpleasantly so. 

If the best results, therefore, are to be hoped for, each 
marking must be observed at the oppositions specially suited to 
it. For as Mars comes to opposition at changing seasons of its 
year, and the conspicuousness of any marking is apparently a 
matter of the seasonal change then in progress upon its hemi- 
sphere, different oppositions are favorable to special features. 

In the subjoined table the figures in the first column repre- 
sent the observed longitudes ; those in the third the longitudes 
reckoned from the Fastigium Aryn, the latter being therefore the 
true Martian longitudes referred to the established zero meridian. 



400 



PERCIVAL LOWELL 



DETERMINATION 


OF POINTS ON THE SURFACE 


OF MARS 


1894. 


No.ofOI». 


Weight 




















Long. 


T^», 


Loag. 


Lat. 


Long. 


L«t. 


LOQff. 


Lbl 








10 




3 


I 


Fastigium Aiyn 


5M 


2-.8 


0% 


N. !•. 






I 






West Sabseus Sinus 

(bottom of gulf) 


I0^4 




5^3 








2 






East cape of Margaritifer 
Sinus 
Margaritifer Sinus 


I5'.5 




I0-.4 








I 






2I'.4 




I6\3 














(mouth of the Indus) 














I 






Margaritifer Sinus 
(mouth of the Hydaspes) 


25*.5 




20-.4 








2 






Aromi Promontoiy 


35M 




30'.o 


S.6*. 






2 






Aurorse Sinus (center) 


54M 




49'. 


S.I1*. 






2 






West Ganges (mouth of) 


59'.7 




54-.6 




, , 




, , 


I 


Lacus Lunse 








N.24*.4 






4 


2 


Solis Lacus (center) 


9i'.3 




86«.2 


S.28V2 






2 


I 


Lacus Phcenicis 


II2°.8 




107*.7 


S. 16%7 


3 


, , 


2 


, , 


Mare Sirenum (beak of) 


I26'.6 












•• 






Mare Sirenum 

(middle of N. coast) 








S.29'.9 


•• 




•• 






Oasis (June. Gigas and 
Pyriphlegethon) 








N.5*. 




, , 




. 




Oasis Qunc. Pyriphlege- 
thon and Steropes) 


I42".7 




I37*.6 
































Oasis (June Eumenides 
and Gigas) 


149'.5 




144*4 


S.7*.t 




, , 




, 




Oasis on Orcus 


I54'.6 




I49*.5 






•• 




• 




Oasis (June. Orcus and 
Steropes) 


l59^2 




I54*.i 














Sinus Titanum 


I74'.5 




i69'.4 


S.20\ 












Scamander 


203'.5 




I98'.4 


S.35".7 




, , 








Trivium Charontis 


208'.5 




203'.4 






•• 




• 




Mare Cimmerium (mouth 
of the Palinurus) 


215'. 




209'.9 






•• 




• 




Mare Cimmerium (mouth 
of the Avemus) 


221 •. 




2I5'.9 






•• 




• 




Mare Cimmerium (mouth 
of the i«:thiops) 


243*. 




237*9 














Eridania (center) 
North coast of Islands 


2X8*. 




2I2'.9 


S.32*.6 
S.33*.6 




, , 




, 




Ausonia (center) 


248-.4 




243*.3 










I 


Lybia (southern 

extremity) 


272'. 




266-.9 


S. 8*.5 








I 


Circe Promontory 


28l*. 




275*.9 


S. 6*.7 




•• 




•• 


Syrtis Major 
(mouth of the Astapus) 


292'.4 




287*.3 






3 


2 • 


2 


Hellas (center of 

northern end) 


303^8 




298'.7 


S.30*.3 




7 




2 


Hammonis Comu 


3I9'.6 




3M*.5 


S.ii'.6 










Euphrates ) mouths 


336M 




331*. 






6 




3 


• of double 


339'.4 




334*.3 


-S.9*.o 




, , 






Phison ) canals 


343*.i 




33«*. 






3 


3 


2 


Edom Promontory 


357'.6 




352..5 


S.8*.2 



Lowell Observatory, 
February, 1895. 



A COMBINATION TELESCOPE AND DOME. 
By A. E. Douglass. 

By a curious coincidence this form of telescope mounting 
was completed in its essentials on the day in which the writer 
first heard of Sir Howard Grubb's ''aquatic" mounting for a 
reflector.' The two mountings have features in common, of 
which the most important is an attempt to procure rigidity of 
support and steadiness of movement. 

Complaints of unsteadiness while using the micrometer are 
common against the ordinary form of mounting. This instabil- 
ity is due to irregularities of the clock movement and lack of 
rigidity in the mounting. A slight wind becomes a most annoy- 
ing visitor and an accidental touch sets the tube vibrating in a 
most aggravating manner. The first idea in the development of 
the spherical telescope (a provisional name adopted here for 
convenience) was that much movement could be eliminated by 
applying the motive power near the eye-end of the tube. Still 
further stability would be insured if the tube were held at each 
end instead 6i at the middle. The final step was to place the 
tube inside a sphere mounted like an ordinary globe. The fol- 
lowing pages will present various mechanical difficulties of the 
plan and suggest solutions. As a matter 6f convenience the 
dimensions recommended will be such as might apply to a sphere 
100 feet in diameter and a lens of seventy -two inches. 

I. SUPPORT AND ADJUSTMENT OF SPHERE. 

The sphere floats on water which is confined in a circular 
cistern of sufficient diameter and depth. At the bottom ^rc 
several supports upon which it can rest while in process of con- 
struction or repair. Its nornflal position would be less than one 
foot above these. The sphere itself should be made of thin steel, 
well braced. It is not necessary that it should have a perfectly 

' In //ew York ffMi/ (about) November ii, 1894. For original article see KhtwI- 
idgilox May, 1894. 

401 



402 



A. E. DOUGLASS 



spherical surface, but it can be put together of flat plates. The 
bearings for the axis consist in anti-friction wheels which are 
mounted in a single casting capable of vertical motion, between 
guides, of about one foot. Fine polar adjustment can be effected 
by some movement within this casting. Each casting is sup- 
ported on a rod which passes into a cylinder below and is 
attached to the center of a transverse diaphragm of slightly flexible 




fvvyr 



Explanation of Fig. i. — A and B are the poles of the sphere. C is the objective 
and D is the eye-end of telescope. F F' is the walk passing beside the zone 
in which revolves the declination carriage. G is the driving motor and machinery. 
H, H, etc., are the supports for dome during construction or repair. I and I show 
the water level for eight-foot draught. J, J are the shutters. K is the tank near 
objective. The entrance will be near F'. 

material dividing it into two parts.' The cylinders are filled with 

water and the upper part of one is joined to the lower part of the 

other, so that by a transmission of pressure from one diaphragm 

to the other any variation in one pole causes an equal variation 

'The idea of usihg a diaphragm instead of an ordinary piston is due to Mr. G. 
Sykes of this town. I have also to thank him for other important suggestions and 
for much assistance with mechanical problems. 



A COMBINATION TELESCOPE AND DOME 403 

in the other. By this arrangement also the sphere sinks deeper 
into the water with added weight, or rises with lessened weight, 
and pressure on the bearings is only momentary. For high lati- 
tudes the plan should be somewhat modified, but I am inclined 
to think that large telescopes of the future, if properly located, 
will be between 15^ and 25° from the equator. 

A brief computation shows that a spherical shell of half-inch 
steel (which I am told is a reasonable allowance for bracing), 
100 feet in diameter, must weigh approximately 300 tons, and 
when floating on water have a draught of eight feet. The addi- 
tion of 300 pounds (one person's weight doubled by the balanc- 
ing system adopted) would cause it to sink -^ inch deeper in 
the water. If the diameter of the cistern is seventy feet and the 
water is allowed to crowd up around the dome its theoretical 
sinkage will be only -^-^ inch. 

II. RIGHT ASCENSION MOVEMENT. 

a. Driving mechanism; slow motions. 

The driving-gear is located within the dome at its equator, 
nearly opposite the opening for the lens. Fig. 2 gives a 
scheme for the driving apparatus, in which B is an electric motor 
whose rate of revolution is controlled by a tuning-fork (see F. L. 
O. Wadsworth in The Astrophysical Journal, February 1895, 
pages 176-7). A is a sprocket wheel which is, in connection with 
an ordinary weight driving-gear, to be used when the electric power 
fails. This wheel is ordinarily loose on its axle, but can be con- 
nected at will. Wheels C and E carry the power to G, where it 
passes a "mouse control" of the Greenwich pattern. From M 
the motion passes through P and Q and the beveled wheels R 
and S. The last mentioned, being the last wheel inside the 
dome, should have an index connected with it which will easily 
give the hour angle within one second of time. The axle of S 
passes through the floor of the dome to the wheel T outside, 
which, engaging with a circular rack, whose radius is slightly 
greater than that of the dome, causes the entire sphere to 
revolve. As the dome is made to sink deeper in the water with 



404 



A. E, DOUGLASS 



increase of weight, the rack must descend with it. The required 
motion can be imparted by using as support to the rack suitable 
bearings placed upon the surface of the sphere, while the rack 
itself is prevented from revolving by stops which slide against 
fixed brackets in the tank. 

b. Rapid motion in right ascension. 
This is obtained by moving C and D bodily towards the 



MOTOR 




Fig. 2 

motor so that C is wholly disengaged from E, and D transmits 
the power to P through F. By this means and by an arrange- 
ment of brushes the mouse control and the tuning-fork are cut 
out during rapid motion. 

c. Backlash absorber. 

T', S', R', Q', P', Y, and X are so geared with respect to 

T. S, R, Q, P, N, and V, that V and X revolve at the same rate 

and in the same direction. If, therefore, by means of a spring, 

W, we produce in V and X a tendency to turn in opposite direc- 



A COMBINATION TELESCOPE AND DOME 405 

tions, the teeth of T and T' will oppose each other by a force 
depending on the strength of W. This will produce an increase 
of friction between T and U, but the amount can be regulated 
by the tension of W. By means of this spring, however, both 
backlash and accidental slipping of T in U may be avoided. The 
displacement of a star in the focus is twice the amount of such 
slipping. 

d. Automatic balance. 

A tank of thirty to sixty cubic feet capacity is placed near the 
objective and moves with it. When the sphere is thrown out of 
balance a change occurs in the otherwise constant strain upon 
each wheel of the driving-gear. The wheel M (Fig. 2) is there- 
fore made in two parts, held in a normal position by a spring, 
but capable of a slight motion with reference to each other in 
either direction. Electrodes are so placed that different circuits 
are closed according to the direction of motion of these parts, 
and an arrangement is connected with the right ascension index 
so that on the meridian these circuits are reversed. One circuit, 
by opening valves, admits water into the tank and the other 
allows water to escape. In this manner and by a proper arrange- 
ment of wires balancing can be effected automatically. To avoid 
possible disturbance during delicate work a switch at the observ- 
ing chair can clamp the two parts of M together. 

III. DECLINATION MOVEMENT. 

The sphere has a long opening in it extending from the upper 
pole along a meridian to within a reasonable distance of the 
opposite horizon. The g^eat circle including this opening is the 
declination circle. The declination axis, in the center of this 
circle and perpendicular to it, is stationary in the sphere. The 
tube passes up to one side of the axis and is supported in a 
special framework which rotates about the axis, the other side of 
the framework, or carriage, being extended for balancing pur- 
poses ; a ladder runs from top to bottom. A strong clamp is 
desirable at each side of the top and bottom to keep the cak-riage 
perfectly steady and aid in rigidity of the dome. By taking the 



406 A. E. DOUGLASS 

best advantage of every opportunity of bracing there seems to be 
no doubt that the two halves can be made sufficiently rigid with 
respect to each other. At the top of the declination carriage is 
the tank previously mentioned. Its best form is tubular and it 
should be placed parallel to the telescope. By making the 
objective project ten or twenty feet and surrounding the upper 
part of the tube by an air space connecting with the interior, the 
size of the sphere may perhaps be appreciably reduced. The 
declination carriage will be rotated by a crank and cog-wheel 
near the observing chair. An index can be placed on the car- 
riage moving along a stationary scale of which each degree will 
be about one foot ; setting within a small part of a degree will 
be easy. 

IV. SUSPENSION OF THE TELESCOPE TUBE. 

The lens is supported in a massive compass-mounting which 
will allow the eye-end of the tube to occupy any part of a circle 
whose diameter is about i°. The chief function of the tube will 
be to preserve the alignment between the objective and eyepiece, 
and it may be made as light as will be consistent with rigidity. 
The position of the eye-end will be controlled by two large 
micrometer screws, whose revolutions may be read off from a 
scale, thus giving us a micrometer of unusually extensive field. 
This plan, or some modification allowing rotation, admits the use 
of high-power negative eyepieces for measuring purposes. A 
small position micrometer should be attached to the draw-tube. 

Finder. 
The finder need not be very large, but should be mounted so 
that its focus would be decidedly below that of the g^eat tele- 
scope. Then by means of two mirrors its optical axis could be 
brought near and parallel to that of the other instrument and its 
focus rendered equally accessible to the observer. A rotating 
disk near its focus would carry eyepieces of different power. 
When an object had been found with a low power and large field 
and brought to the intersection of cross-threads, a high power 
could be turned in at once which would insure its visibility in the 



A COMB/NATION TELESCOPE AND DOME 407 

large instrument. A rotary movement in either one of the elbows 
will throw the eye-end of the finder out of the way when the 
object is finally in position. 

V. SHUTTERS AND CAP. 

The long portions of the opening which are not covered by 
^ the top of the declination carriage are closed by automatic shut- 

ters which open on the approach of the lens and close on its 
retreat. This is effected by a system of cog-wheels and eccentrics. 
;^ It will probably be found best to extend the top of the carriage 

:r in either direction by a light screen to prevent too free entrance 

r: of outside air. The entrance of cold night air might be regulated 

7;} by ventilators capable of adjustment. 

The cap for the lens consists of two halves which separate in 
opposite directions to balance each other, controlled by a wind- 
^^ lass near the observing chair. 

'\\' VI. OBSERVING CHAIR AND ACCESSORIES. 

K'> An iron post projects from the floor of the declination car- 

:(- riage in line with the optical axis of the telescope in its central 

'<: position. About this as an axis a car revolves on a circular 

;i^' track, supporting the observing chair proper on radial tracks. 

;:^ The chair may be constructed to suit individual taste ; the fol- 

>^< lowing plan will serve as an example. Near the center of the chair- 

^>.: base two boards are hinged, one supporting the back and head, 

^< and the other the knees and feet. The latter has a second hinge 

^.-^ at the knees, so that by drawing the footboard from its outer- 

most position towards the center a seat is formed (for observa- 
tions at low altitudes). The headboard should be kept raised 
y::'^ a little and made adjustable through a small arc, and the whole 

the^'' cushioned. At the top of the headboard is a transverse strip of 

wood which can be raised or lowered through the space of two 
inches by a lever which passes down beside the headboard to be 
within reach of the hand. Upon this strip a small head-rest is 
placed having a lateral sliding motion. On either end of the rest 
is a mirror placed at an angle of 45'', by means of which the 
micrometer readings may be taken; otherwise the eye can 



408 A. E. DOUGLASS 

scarcely get far enough away to see the scale. Eyepieces 
should be kept in a drawer in one side of the chair so that they 
can be reached without getting up. A switchboard containing 
all necessary electric connections can be suspended by the slack 
of its wires from some point above so that it can be hooked on 
to the observing chair within reach. 

Time. 
Time had best be brought in to a sounder from a standard 
clock in an adjoining building. This may be used to correct a 
watch mounted, I would suggest, on the telescope itself, facing 
the observer, so that the time may be noted as soon as an 
observation is complete. 

Recording, 

It might be possible to attach to the telescope or head-rest 
two mirrors which would bring the record-book into view with 
one eye while the other was occupied at the eyepiece. This 
would be especially serviceable in making drawings of planetary 
detail. 

VII. ENTRANCE. 

The outside door is in the declination circle, and as far as 
possible below the lower pole without interfering with the motion 
of the carriage. The door itself is arranged to rotate about its 
center so that in any position of the dome it can be turned 
upright. It is reached by a U-shaped stairway. Inside, from 
the level of the door throughout the whole range of movement 
of the lower end of the carriage, a pathway, a quadrant in section, 
is built on each side of the declination zone. The two pathways 
if placed together would make a semicircle with its center 
tangent to the inner surface of the sphere. When the hour angle 
is zero a person may step from the carriage to the pathway on 
either side, the nearest part of it being then level ; for eastern 
hour angles, he can find on the eastern pathway a part where the 
east and west inclination will be very small ; for western hour 
angles he will turn to the western pathway. The radius of 
curvature of the section of these pathways could be perhaps five 



A COMBINATION TELESCOPE AND DOME 409 

feet, and the ways themselves could be divided into separate 
paths by raised strips so as to make it easier to walk on them. 
Each of these separate paths will be supplied with steps of the 
proper height or with cleats nailed across, as best suits the inclina- 
tion at which it will be used. At the equator a side path will 
extend at right angles to give access to the motors, driving-gear 
and desk. One-half of this path will have cleats across for small 
inclinations of the dome and the other half will be a stairway 
which can be used for large inclination in either direction. The 
desk will be of a form suited to the necessities of the case. 
The stand in front of it and its own top will both be curved 
surfaces so that one may always find a level standing place and 
a level part of the desk to write upon. These two surfaces will 
be portions of concentric cylinders whose axis is parallel to the 
axis of the dome. 

A second entrance is placed near the objective, with a ladder 
bolted to the outside of the sphere for approaching it. 

VIII. COMPARISON WITH OTHER MOUNTINGS. 

The forms of equatorial mountings now in use differ as to 
their stability ; perhaps the best in this respect is the Equatorial 
Coudi. By shortening the equatorial arm increased firmness will 
result, but it always has a disadvantage in requiring two mirrors. 
Irregularities in clock motion may occur in the Coudi as well as 
in any other form, whereas with a spherical telescope, the power 
being applied at the greatest possible distance from the axis, 
imperfections in the clock appear unchanged instead of highly 
magnified. Its great inertia also will aid in giving a constant 
motion. Apparently the individual feature in this mounting 
which cannot be duplicated in any other is the application of 
clock-power so far from the axis, resulting in great strength and 
steadiness. 

In this connection it will be interesting to note what power 
is now in use in telescopes of some size — remembering that the 
single right ascension movement of the design under considera- 
tion includes both right ascension movement and turning of 



410 



A. E. DOUGLASS 



domes in the ordinary mounting. Through the courtesy of the 
Director and Secretary of the Lick Observatory, and of Messrs. 
Warner & Swasey, I have received data on this subject which 
I can here present : 



RIGHT ASCENSION MOVEMENT. 



TeI«Mope 


Size 
in. 


We'«ht 
niovcd 
toot 


Skm 

motkm 

H.P. 


Qukk 

motion 

H.P. 


Time 

I Rot. 

m. 


Work 
ft.lU. 


Maker 


Lowell... 
Naval Ob. 
Lick .... 
Yerkes... 


I8 
26 
36 
40 


about I 
14.2 


0.0003 
0.0020 

0.0033 
0.0107 


0.0455 

0.0295 
O.II75 


0.5 

3.2 

4.0 


750 

3.II0 
15.510 


Clark 

Warner & Swaser 
44 

44 



POWER REQUIREEt FOR TURNING DOMES. 



Dome 


Diam. 
ft. 


Weight 
tons 


Power 


Time 

X Rot. 

m. 


Work 
ft. lbs. 


Designer 


Lowell'... 
Naval Ob.. 

Lick 

Yerkes.... 


35 

45 

75 
90 


3 

100 
165 


0.22 
0.07 
1. 18 


2.0 


14,400 

5.773 

312,000 


W. H. Pickering 

Warner & Swasey 

Union Iron Works, San Francisco 

Warner & Swasey 



From an examination of the above figures it seems probable 
that one horse-power is not an over-estimate for a telescope of 
seventy-two inches aperture. If we could entirely disregard 
friction this would undoubtedly be far more than sufficient ; but 
in a new machine friction is a very uncertain factor as well as a 
most important one, and actual experiments are the only reliable 
source of information. The only satisfactory form that such 
experiments could take would be to try the dome on a small 
scale, as, for instance, arranged for an eighteen-inch telescope. 
A study of the different sources of friction leads me to think 
that the mechanical power required will be small. 

Lowell Observatory, 
March. 7, 1895. 

* The design of this dome closely follows that suggested in A. ami A,^ January, 
1894. The work of turning it could be enormously reduced by substituting iron for 
wooden tracks and stiffening the live-ring horizontally. 



STARS HAVING PECULIAR SPECTRA.' 



ELEVEN NEW VARIABLE STARS. 



By M. Fleming. 

An examination of the photographs of stellar spectra 
taken at Arequipa, Peru, under the direction of Professor S. I. 
Bailey, and forming part of the work of the Henry Draper 
Memorial, has, during the last few months, shown marked pecu- 
liarities in the spectra of eleven objects enumerated in Table I, 
and has resulted in the discovery of eleven new variable stars 
enumerated in Table II. The first column of Table I gives the 
designation of the object and is followed by the approximate 
right ascension and declination for 1900, the catalogue magni- 
tude, and a brief description of the photographic spectrum : 







TABLE I. 






Desifiiatioii 


R.A. 

Z900 


Dec 
1900 


Mag. 


DeKripckm 


Z,C. 3*» 1404 

S.B J?,— 22^* 1070 

//.G,C, 2070 
A.G.C. 8518 
A,G.C. 9181 


3N6-.7 

5 14 .5 

5 40 .5 

6 46 .1 

7 10 .2 
7 19 .5 
7 41 I 
7 42 .0 

16 16 .8 

16 39 .8 

17 13 .2 


— 43" SO 

— 22 19 

— 69 43 

— 32 24 

— 26 10 
+ 9 7 

— 31 -41 
-34 8 

— 51 18 
-67 36 

— 66 IS 


8.5 
8.7 

4.0 
5.4 

9.2 
9.5 

8.5 


Type IV 

Peculiar 

Bright lines, Gas. Neb. 

J/fi bright 

J/fi bright 

Bright lines. Gas. Neb. 

TypeV 

Bright lines, Gas. Neb. 

TypeV 

Type IV 

Peculiar 


C0rd.D,— 3i'' S004 
Z,C 7^ 2999 




^C. I7»» 734 



The gaseous nebula whose approximate position for 1900 is 



in R.A. 



7»» 19" 



,5, Dec. + 9^ 7' was found in the examination of 



the photographs taken with the 8-inch Draper telescope at Cam- 
bridge. 

' Communicated by Edward C. Pickering, Director of the Harvard College Obsenr- 
ator\'. 

4" 



412 



if/. FLEMING 



The stars contained in Table II have a spectrum of the third 
type, having also the hydrogen lines bright, and their variability 
was at once suspected from this peculiarity. Conclusive evi- 
dence of their variation, as shown below, was in each case 
obtained on examination of chart plates of these regions. They 
are not here announced as suspected variables, but as variable 
stars, the variation, in each case, having been confirmed inde- 
pendently from an examination of the photographs by Professor 
Edward C. Pickering. The first column gives the constellation, 
and the second, the catalogue designation. This is followed by 
the approximate right ascension and declination for 1900, the 
number of photographs examined, and the maximum and mini- 
mum photographic magnitude as derived from the photographs : 

TABLE II. 



ConsteU. 



Tttcana 

Pictor 

Lepus 

Pictor 

Scorpius 

Teiescopium 

Indus 

Octans 

Gnis 

Aquarius 

Phcenix 



Designation 



A,G,C, 5428 
5.^./?.— 22° 995 
ZC. ^ 283 



Z,C, 20»» 1539 



A.G,C, 32334 



R.A. 
X900 



0»» l8".4 

4 43 .5 

5 o .6 

5 8 .3 

17 35 .1 

20 7 .6 

20 49 .0 

20 57 .4 

21 42 .1 

22 13 .2 

23 S3 .9 



Dec 
xgoo 



-62'' 14 

-49 25 

-2Z 2 

-48 38 

-43 42 

-47 18 

-54 42 

-82 30 

-47 22 

-21 26 

-57 8 



No. Plates 



15 
16 

13 
16 

37 
7 

26 
28 

25 
18 

22 



Mag. 



Max. 



8.7 
8.1 
8.2 
8.6 

9.3 
8.4 
8.4 
9.0 
8.4 
8^ 
7.2 



MIn. 



<ii.6 

9.5 
10.9 

<I3.3 
12.7 
1 1.6 

<I2^ 
<I2.5 

<I2.5 

1 1.6 
8.7 



— Tucanae, R.A. o** i8°.4, Dec. —62'' 14'. The magnitudes 
of this star as derived from photographs taken on September 1 1 , 
October 8, November 28, 1889; September 12, September 23, 
1890; August 20, September 17, October 4, October 4, October 
26, 1891; July 31, September 5, 1892; October 23, 1893; July 
6, and July 23, 1894, are <io.9, 10.6, 9.4; 9.4, 9.7; <io.7, 
<ii.3, <ii.s, <ii/6, 1 1.3; 8.7, 9.2; <io.3; 8.8 and 9.0 
respectively. 

— Pictoris, A.G.C. 5428. The magnitudes of this star as 



STARS HA VING PECULIAR SPECTRA 4 1 3 

derived from photographs taken on September 26, September 
26, October 4, October 8, 1889; September 23, 1890; October 
26, 1891 ; September 28, September 29, December 8, 1892; 
November 18, 1893 ; August 21, August 25, September 14, Sep- 
tember 19, September 19, and October 19, 1894, are 8.5, 8.3, 8.5, 
8.6; 8.8; 8.6; 8.6, 8.6, 9.5; 9.0; 8.1, 8.2, 8.4, 8.5, 8.5, and 8.9 
respectively. 

— Leporis, S,B,D, — 22^995. The magnitudes of this star 
as derived from photographs taken on November 4, November 

17, 1889; February 3, September 19, September 19, December 
29, 1890; January 26, 1891; November 18, 1893; September 
14, October 24, 1894; March 5, March 9, and March 16, 1895. 
are 9.2, 8.4; 9.4, 10.5, 10.3, 8.4; 9.4; 8.2; 10.4, 9-0 ; 10.9, 10.8, 
and 10.8 respectively. 

— Pictoris, Z,C. 5** 283. The magnitudes of this star as 
derived from photographs taken on September 26, September 
26, October 8, November 6, 1889; September 23, 1890; Octo- 
ber 24, November 23, 1891 ; September 28, October 6, Decem- 
ber 8, 1892; September 27, September 27, November 18, 1893 ; 
September 14, September 19, and November 5, 1894, are <io.8, 
<ii.i, <ii.8, <I2.7; <I3.2; <I3.3» <ii.4; U-S, 12.0, 
<I3.3; 10.6, 10.6, 11.4; II. 3, 1 1. o, and 8.6 respectively. 

— Scorpii. R.A. .17** 35".i, Dec. —43** 42'. The magnitudes 
of this star as derived from photographs taken on July 9, July 9, 
July 13, July 17, July 19, July 20, July 21, July 22, August 6, 
August 28, 1889; May 9, June 9, June 9, 1890; May 18, May 

18, May 19, May 19, September 8, 1891 ; April 19, June 6. June 
10, 1892 ; April 27, May i, May i, May 8, May 8, June 24, 1893 ; 
May 23, June i, June i, June i, June i, June 14, July 20, July 20, 
August 14 and September 21, 1894, are 9.9, 9.8, 9.8, 9.6, 9.8, 
9-6, 9-5. 9-5. 9-3. 9-6; ii-4, 10.7, 10.8; <ii.8. <io.7, 12.4, 
<I2.2, 9.6; <ii.i, 10.8, 10.8; <ii.i, 12.6, <ii.6. 12.7, 
<I2.4, II. 5; <I2.3, <i2.i, <ii.8, <ii.6, <ii.3, <I2.3, 
1 1.4, 1 1.2, 10.7 and 9.9 respectively. 

— Telescopii. R.A. 20** 7".6, Dec. -47° 18'. The magnitudes 
of this star as derived from photographs taken on July 24, 1893 ; 



414 M. FLEMING 

July 21, August 21, September 12, September 15, September 15 
and September 27, 1894, are 8.7 ; 1 1.6, 1 1.6, 9.0, 9.1, 8.9 and 8.4 
respectively. 

— Indi. Z.C 20** 1539. The magnitudes of this star as 
derived from photographs taken on June 20, July 5, July 19, July 
21, July 22, August 22, October 26, 1889; May 21, May 22, 
June 2, June 10, June 10, July 16, 1891 ; September 8, September 
8, September 8, 1892; June 27, July 31, July 31, September 23, 
September 23, 1893; May 21, July 13, July 25, August 11 and 
August II, 1894, are 10.3?, <ii.3, <I2.3, <ii.9, <ii.6, 
<I2.4, <9.9; II. 3, 11.7, <I2.I, <io.4. 12.0, <io.2; <ii.i. 
<I2.2, <I2.2; 10.2, 11.3, 11.4, <I2.2. <I2.3; 8.4, 9-3. 9-7, 

1 0.1 and 1 0.1 respectively. 

— Octantis. R.A. 20** 57".4, Dec. —82** 30'. The magnitudes 
of this star as derived from photographs taken on June 1 7, June 
18, June 20, August 8, September 4, September 16, September 
27, October i, 1889; June 9, June 13, June 14, August 5, Sep- 
tember 5, 1890; June II, June 11, June il, June 14, June 19, 
October 19, October 19, 1891 ; May 3, June 23, 1893; July 16, 
July 16, August 2, August 13, August 14 and September 10, 1894, 
are 12.4, 11.8, < 1 1.7, < 10.8, < 10.4, < 10.3, < 12.3, <i 1.8; 9.8, 
9.5,9.5, 10.3, 12.0; <ii.2, <io.4, <I2.5, <io.9, <I2.5, 10.6. 
10.6; <I2.5, 9.8; 1 1.3, 1 1.2, 9.0, 9.1, 9.0 and 9.1 respectively. 

— Gruis. R.A. 21** 42". i, Dec. —47*' 22'. The magnitudes of 
this star as derived from photographs taken on June 20, June 20, 
July 13, July 13, July 13, July 18, July 18, July 22, September 
II, September 28, October 8, 1889; June 12, 1890; May 20, 
June 8, June 8, June 8, June 13, July 16, 1891 ; October 6, 
October 6, 1892 ; July 24, July 24, August 21, August 21, 1893 ; 
May 21 and August 31, 1894, are <io.5, <io.5, <ii.i, <ii.5, 
<ii.7, <I2.4, <I2.5, <i2.3; <ii.6, <io.5, <ii.7; 11.9; 
<ii.8, <io.5, <ii.8, <io.i, <io.8, <I2.2; 8.8, 8.8; 11.8, 
1 1.7, 10.2, 10.6; <ii.6 and 8.4 respectively. The magnitude 

10.2 on August 21, 1893, may be somewhat in error since the 
image of the variable is near the edge of the plate, thus render- 
ing the comparison difficult. 



STARS HA VING PECULIAR SPECTRA 4 1 5 

— Aquarii. R.A. 22** I3".2, Dec. —21** 26'. The magnitudes 
of this star as derived from photographs t^ken on August 28, 
September 3, September 25, September 27, September 30, Sep- 
tember 30, October 24, 1889; July 7, July 10, 1890; June 14, 
June 16, 1891; July 3, 1892; September 20, September 20, 
October 25, October 25, 1893; August 10 and August 11, 1894, 
are lo.o, lo.o, <io.6, 10.6, 10.8, 10.8, 11.6; lo.o, 9.9; 10.9, 
<io.6; <io.4; 9.4, 9.4, 9.3, 9.3; 8.5 and 8.4 respectively. 

— Phoenicis. A,G.C, 32334. The magnitudes of this star as 
derived from photographs taken on July 17, August 20, Sep- 
tember II, October 8, October 8, 1889; June 30, August 20, 
August 20, August 21, September 17, 1891 ; May 16, May 16, 
August 16, September 5, 1892; July 24, July 27, August 21, 
September 20, September 27, October 23, November 17, 1 893; 
and July 24, 1894. are 8.2, 8.4, 8.4, 7-2, 7-5; 8.7, 7.8, 7.7, 8.2, 
8.3; 8.2, 8.2, 8.4, 8.4; 8.4, 8.2, 8.2, 8.i» 8.3, 8.2, 8.6; and 8.0 
respectively. 

The magnitudes given above for the new variable star in 
Aquarius were derived from the mean of two or more measures 
made independently on two different dates, and using a different 
set of comparison stars when the variable was faint. The 
average difference of these measures is + .09. 

Harvard College Observatory, 
Cambridge, Mass., April 9, 1895. 



A SPECTROSCOPIC PROOF OF THE METEORIC CON- 
STITUTION OF SATURN'S RINGS. 

By James E. Keeler. 

The hypothesis that the rings of Saturn are composed of an 
immense multitude of comparatively small bodies, revolving 
around Saturn in circular orbits, has been firmly established 
since the publication of Maxwell's classical paper in 1859. The 
grounds on which the hypothesis is based are too well known to 
require special mention. All the observed phenomena of the 
rings are naturally and completely explained by it, and mathe- 
matical investigation shows that a solid or fluid ring could not 
exist under the circumstances in which the actual ring is 
placed. 

I have recently obtained a spectroscopic proof of the meteoric 
constitution of the ring, which is of interest because it is the first 
direct proof of the correctness of the accepted hyp>othesis, and 
because it illustrates in a very beautiful manner (as I think) 
the fruitfulness of Doppler's principle, and the value of the spec- 
troscope as an instrument for the measurement of celestial 
motions. 

Since the relative velocities of different parts of the ring 
would be essentially different under the two hypotheses of rigid 
structure and meteoric constitution, it is possible to distinguish 
between these hypotheses by measuring the motion of differ- 
ent parts of the ring in the line of sight. The only difficulty 
is to find a method so delicate that the very small differences of 
velocity in question may not be masked by instrumental errors. 
Success in visual observations of the spectrum is hardly to be 
expected. 

Soon after the large spectroscope of the Allegheny Observa- 
tory was completed, in 1893, ^ attempted to determine the rela- 
tive motions of different parts of the system of Saturn, by photo- 
graphing the spectrum with the slit parallel to the major axis of 

416 



METEORIC CONSTITUTION OF SA TURN'S RINGS 4 1 7 

the ring, but failed to obtain satisfactory results. The unfavora- 
ble atmospheric conditions at Allegheny, the strong yellow color 
of the objective of the thirteen-inch equatorial, and the yellow 
color of Saturn itself so reduced the intensity of the violet part 
of the spectrum that the negatives obtained with a sufficiently 
high dispersion were too weak and granular to admit of measure- 
ment. Another unfavorable circumstance was the fact that I had 
to guide the practically invisible image corresponding to the Hy 
line by means of the visual image, which was greatly out of focus 
on account of the chromatic aberration of the visually corrected 
telescope. Having recently obtained excellent results in other 
directions with orthochromatic plates, by the use of which the 
difRculties mentioned above are to a great extent obviated, I was 
induced to repeat my earlier attempts, and obtained two fine pho- 
tographs of the lower spectrum of Saturn on April 9 and 10 of 
the present year. The exposure in each case was two hours, and 
the image of the planet was kept very accurately central on the 
slit-plate.. After the exposure the spectrum of the Moon was 
photographed on each side of the spectrum of Saturn, and nearly 
in contact with it. Each part of the lunar spectrum has a width 
of about one millimeter, which is also nearly the extreme width 
of the planetary spectrum. On both sides of the spectrum of 
the ball of the planet are the narrow spectra of the ansae of 
the ring. 

The length of the spectrum from ^ to D is 23 millimeters. 
The focus was adjusted on the line A 5352, a little above the posi- 
tion of maximum sensitiveness of an orthochromatic plate, in the 
yellow green. On both plates the densities of the different spec- 
tra are very nearly equal, and the definition is excellent. It is 
hardly necessary to say that all the lenses used in the apparatus 
are visually corrected objectives. 

These photographs not only show very clearly the relative 
displacement of the lines in the spectrum of the ring, due to the 
opposite motions of the ansae, but exhibit another peculiarity, 
which is of special importance in connection with the subject of 
the present paper. The planetary lines are strongly inclined, in 



4 1 8 JAMES E. KEELER 

consequence of the rotation of the ball, but the lines in the spec- 
tra* of the ansa: do not follow the direction of the lines in the 
central spectrum; they are nearly parallel to the lines of the 
comparison spectrum, and, in fact, as compared with the lines of 
the ball, have a slight tendency to incline in the opposite direc- 
tion. Hence the outer ends of these lines are less displaced than 
the inner ends. Now it is evident that if the ring rotated as a 
whole the velocity of the outer edge would exceed that of the 
inner edge, and the lines of the ansae would be inclined in the 
same direction as those of the ball of the planet. If, on the other 
hand, the ring is an aggregation of satellites revolving around 
Saturn, the velocity would be greatest at the inner edge, and the 
inclination of lines in the spectra of the ansae would be reversed. 
The photographs are therefore a direct proof of the approximate 
correctness of the latter supposition. 

To apply more precise reasoning to the subject under consid- 
eration, let us determine the form of a line in the spectrum of 
Saturn when the slit is in the major axis of the ring, on the assump- 
tion that the planet rotates as a solid body and the ring is a swarm 
of particles revolving in circular orbits according to Kepler's 
third law. At present the motion of the system as a whole is 
neglected. The upper part of Fig. i represents the image of 
Saturn on the slit of the spectroscope (the scale above it applies 
to the instrument used at Allegheny), and the narrow horizontal 
line in the lower part of the figure represents an undisplaced line 
in the spectrum, or solar line.' Let this line be taken as the axis 
of jr, and the perpendicular line through its center as the axis of >'. 
The red end of the spectrum is supposed to be in the direction 
of the positive axis of y^ and the camera and collimator of the 
spectroscope are assumed to have the same focal leng^, so that 
the breadth of the spectrum is equal to the length of the illumi- 
nated part of the slit. Corresponding points in the slit and spec- 
tral line will then have the same value of x. 

Now let x,y, be the codrdinates of a point on the displaced 
line, 

■ The curvature of the line in a prismatic spectrum need not be considered. 



METEORIC CONSTITUTION OF SATURN'S RINGS 419 



J L 



One milHmtttr, 

i I I 




w 



Fig. I 



420 JAMES E. KEELER 

z;= velocity of point corresponding to x^y in the line of sight, 
F'=velocity of a point on the equator of Saturn, 
a=angle between the line of sight and the radius of Saturn 

which passes through the point corresponding to x^ y^ 
2p=: width of spectrum, 

/3= elevation of Earth (and Sun) above the plane of the ring.' 
The displacement y is proportional to the velocity in the line 
of sight. Then we have 

:r:=psina, 
y^=^av^aV*^\Vio, cos A 

y aV* 

- = cos a = constant. 

X p 

Hence the planetary line is straight, but inclined to the solar 
line at an angle y, 

^ = tan"' cos /8. 

P 
To determine the form of a line in the spectrum of the ring, 

regarded as a collection of satellites, we have, by Kepler's third 

or, since TV=2irRt 

K- — ^. 
cR 

Since x is proportional to R and j' to v (where «;= velocity in 

the line of sight=Fcos/8), we may write 

xf=b, 

which is the equation to the curve of which the lines in the 

spectrum of the ring are a part. The curve is represented by 

the dotted line in the figure ; it is symmetrical with respect to 

the axis of x, but only the upper branch has a physical meaning, 

and the curve corresponding to the other half of the image is 

obtained by taking both x and y with negative values. 

k 

In the equation F= --7=^, log k — 3.7992 for the Saturnian 
V R 

system, R being expressed in kilometers and Fin kilometers per 

■ The slight error introduced by the assumption that the Earth and Sun are in the 
same direction from Saturn is inappreciable, when Saturn is anywhere near opposition. 



METEORIC CONSTITUTION OF SATURN'S RINGS 4^1 

second. The computed motions of di£Eerent parts of the system 
are given in the following table. The g^uze ring is not con- 
sidered, as its spectrum does not appear on the photographs ; 
the rings A and B are not separately distinguishable. 



Object 


R 


Period of a 
Satellite at 
Distance R 


Velociiy 


Velocity in 
Uaeof Si>ht 
April xo, x89S 


Outer edge of ring 
Middle of ring 
Inner edge of ring 
Limb of planet 


Kilometen 

135.100 

112,500 

89,870 

60,340 


Houn 
13.77 
10.46 

7.47 
4.1 1 


KUonecen 
17.14 
18.78 
21.01 
25.64 


Kilometen 
16.35 
17.91 
20.04 
24.46 


Limb of planet - 


60,340 


Rotation 
l0»».23{A.Hall) 


10.29 


9.82 



With the values given in the above table, and others which 
do not correspond to actual points in the system, the dotted 
curves were platted. For the ordinates, however, twice the 
values in the last column were taken, since the displacement of a 
line, due to motion in the line of sight, is doubled in a case of 
a body which shines by reflected and not by inherent light, pro* 
vided (as in this case) the Sun and the Earth are in sensibly the 
same direction from the body. The planetary line is drawn to 
the same scale, and the heavy lines in the figure represent accu- 
rately the aspect of a line in the spectrum of Saturn, with the 
slit in the axis of the ring, as photographed with a spectroscope 
having about three times the dispersion of my own instrument. 

The width of slit which I used (o"*".028, or 7900^" on the 
surface of Saturn) is also represented in the figure. 

If the whole system has a motion in the line of sight, the 
lines in the figure will be displaced toward the top or the bottom, 
as the case may be, but their relative positions will not be altered. 

It is evident that in making a photograph of this kind the 
image must be kept very accurately in the same position on the 
slit plate, as otherwise the form.of the lines shown in the figure 
would be lost by the superposition of points having different 
velocities. The second plate was made with special care, and as 
the air was steadier than on the first occasion, the definition is on 



422 JAMES E, KEELER 

the whole somewhat better than that of plate i, although the 
difference is not great. On both plates the aspect of the spec- 
trum is closely in accordance with that indidated by theory and 
represented in the figure. The planetary lines are inclined from 
3^ to 4°, and the lines in the spectra of the ansae have the 
appearance already described. The slight curvature of the 
latter indicated by theory is, of course, unrecognizable. On 
account of the extreme narrowness of the spectra (barely more 
than a tenth of a millimeter) it was useless to attempt anything 
like a measurement of the inclination of the lines. The diiec- 
tion of such short lines is frequently masked by irregularities in 
the grain of the plate, and occasionally a line is considerably 
distorted. However, in fifty of the sharpest lines, in the region 
of best definition, only five were inclined in the same direction 
as the lines of the ball, while the rest were inclined as required 
by the theory, or elsewhere apparently parallel to the undisplaced 
lines of the lunar spectrum. 

If the ring revolved as a whole, the displacement of lines in 
its spectrum would follow the same law as for a rotating spheie ; 
that is, the lines would be straight and inclined, their direction 
passing through the origin. If the ring rotated in the period 
of its mean radius, a glance at the figure shows that the lines 
would practically be continuations of the planetary lines. Such 
an aspect of the lines as this would be recognizable on my 
photographs at a glance. 

It will be seen from the foregoing considerations that the 
photographs prove not only that the velocity of the inner edge 
of Saturn's ring exceeds the velocity of the outer edge, but that, 
within the limits of error of the method, the relative velocities 
at different parts are such as to satisfy Kepler's third law. 

Besides (i) the proof of the meteoric constitution of the 
rings, explained above, each line of the photographs gives (2) the 
period of rotation of the planet, (3) the mean period of the 
rings, (4) the motion of the whole system in the line of sight. I 
have measured a number of lines on each plate and compared the 
results with the computed values of the corresponding quantities. 



METEORIC CONSTITUTION OF SATURN'S RINGS 423 

The most accurate method' of measuring the relative dis- 
placement of the opposite ends of a line in the spectrum of the 
planet is to measure the angle ^. 

The value of ^ depends upon the dispersion and oth^r con- 
stants of the spectroscope employed, as well as upon quantities 
which are independent of the instrument. If we let L = the 
velocity of light in kilometers per second = 299,860 ; X = the 
wave-length of the measured line in tenth-meters ; D = the linear 
dispersion of the photographed spectrum at the position of the 
same line, expressed in tenth-meters per millimeter ; p = half the 
width of the spectrum in millimeters ; — we have by Doppler's 
principle, allowing for the double effect already mentioned, 

or, ^ 

iK 

from which we obtain the velocity in the line of sight at the 
limb (F'cos /8) by placing x — p. That is, 

y»^ pZ>Z tan ^ 
iK cos )9 

The value of p is computed from the angular semi-diameter 
of the planet and the focal length of the telescope. It cannot 
be obtained accurately by measurement of the photograph, 
because the borders of the spectrum are indistinct. For my 
instrument at the time of observation, p = o""'.2i34. D is 
obtained from a wave-length curve constructed from measure- 
ments of a standard plate of the solar spectrum made with the 
same apparatus, and ^ is directly measured under a microscope 
provided with a position circle. 

The relative displacement of a line in the spectra of the ansae 
is measured directly, the micrometer wire having first been 

■ This method w due to Deslandres (C. R, xao, 417), and I have found it to be 
very satisfactory. The conclusions in Deslandres' article, with respect to the motion in 
the line of sight of bodies which are not self-luminous, are not new, although they are 
treated more fully than elsewhere. 



424 



JAMES E. KEELER 



placed parallel to the lines of the comparison spectrum. If S is 
this measured interval, the mean velocity of the ring is 

4X cos^ 

The displacement could also be determined by measuring the 
angle which the line joining the centers of the two short lines of 
the ansae makes with the comparison lines. I have found the 
direct method to be preferable. 

There remains the motion of the whole system in the line of 
sight, which has hitherto not been considered. It is best deter- 
mined by comparing the mean of the positions of the lines in 
the spectra of the ansae with the corresponding line of the com* 
parison spectrum. The results for this motion are unsatisfactory, 
as might be expected from the circumstances of observation. 
Owing to the fall of temperature during the rather long exposure 
of two hours, and the fact that the lunar sf>ectrum was photo* 
graphed at the end, and not in the middle of the exposure to the 
planet, the two spectra are relatively displaced by an amount 
which is about ten kilometers greater than that due to the 
motion of Saturn in the line of sight. I have therefore made no 
careful measurements of this displacement. For the reasons 
given above, the planetary lines are somewhat less sharp than 
the lines in the lunar spectrum, which was photographed with an 
exposure of only six minutes. 

The results of all the measurements are given in the follow- 
ing tables : 







Photograph No. 


[, April 9 


. 1895. 






X 


D 


♦ 


^ISJ" 


C—0 


a 


Vegof 


C—0 


Teath- 
meten 

5324.3 
5328.4 
5371.6 
5383.5 
5429.9 


Tenth- 
meten 

27.55 
27.65 
28.77 
29.09 
30.37 




3 36 

3 II 
3 20 

3 8 


Kllometen 
10.92 
13.39 
9.99 
10.56 
10.27 


Kiloneien 
—0.63 
—3-10 
-fo.30 

.-0.27 
+0.02 


0.0456 
0.0464 
0.0404 
0.0362 
0.0402 


18.54 
18.9a 
17.OX 
15.37 
17.67 


Kflo- 

+0.24 
-0.14 
+1.77 
+3.41 
+1.11 








11.03 


—0.74 




17.50 


+ 1.28 



METEORIC CONSTITUTION OF SA TURN'S RINGS 4^5 



Photograph No. 2, April io» 1895. 



X 


D 


♦ 


VelodtTof 

Limb 


C—0 


a 


Mem 

Velocity of 
Ring 


C—O 


Tratn* 


Tenth- 







KUometen 


MiUimelen 


Kikmeicn 


Kilo- 


Bidcn 


meten 












neten 


5324.3 


27.55 


2 II 


6.62 




-3.67 


0.0468 


19.03 


-0.25 


5328.4 


27.65 


3 19 


10.09 


- -0.20 1 


0.0412 


16.81 


--I.97 


5371.6 


28.77 


2 42 


8.47 




-1.82 


0.0436 


18.35 


--O.43 


5383.5 


29.09 


3 13 


10.19 




-O.IO 


0.0420 


17.84 


--O.94 


5429.9 


30.37 


3 49 


12.51 


— 2.22 


0.0468 


20.56 


-1.78 








9.58 


+0.71 




18.52 


+0.26 



The results from both photographs are 

Velocity of limb =10.3 di 0.4 kilometers, 
Mean velocity of ring = 18.0 ± 0.3 kilometers ; 

the computed values being 10.29 and 18.78 kilometers respec- 
tively. 

Although there seems to be no systematic difference between 
the two plates, the results for each differ by more than the prob- 
able error. With photographs on so small a scale, distortions 
of the lines are produced by the irregular deposit of even a few 
particles of silver ; hence it is advisable to measure a large num- 
ber of lines instead of multiplying observations on a few of 
them. 

The number of lines in the table is however sufficient for the 
present purpose. 

As I have already pointed out, it is necessary to guide the 
telescope with extreme accuracy in making such photographs as 
those described in the present paper, and the method which I have 
used is so simple and effective that a short account of it may be 
of interest. 

The spectroscope is fully described in Astronomy and Astro- 
physics, 12, 40, January, 1893, and the prism-train used in these 
observations is there shown in Plate VII. The slit is observed 
during an exposure by a small '' broken " telescope, which receives 
the rays reflected from the first surface of the prism nearest to 
the collimator. 



426 JAMES E. KEELER 

To prepare for an observation of Saturn, the slit is shortened 
until its length is equal to the computed length of the image 
(major axis of the ring). A small bar, which is wider at one 
end than at the other, is cut out of thin metal, and placed across 
the field of the diagonal telescope. If the bar is approximately 
of the right width, then, by throwing the image of the slit a little 
above or below the center, and by rotating the eyepiece, which 
carries the bar with it, the bar can be made to very nearly cover 
the image, leaving a very short length of slit uncovered at each 
end. When the telescope is directed to Saturn the extreme 
ends of the ring appear from behind the (invisible) bar as two 
minute points or stars, and the attention of the observer is 
concentrated on keeping these stars equally bright. Any dis- 
placement in declination is indicated by their disappearance or 
unusual faintness. The photographs show that the guiding 
by this method is quite accurate. The spectra of the ansae 
do not show any traces of the Cassini division, but it would 
probably be requiring too much to expect that they should do 
so, considering the small size of the image and the length of 
the exposure. 

It is a question whether these observations could be better 
made with a larger telescope. If the same spectroscope were 
mounted on a large telescope, the width of the photographed 
spectrum would be greater, the lines would be longer, and their 
direction could be more definitely measured. On the other 
hand, the inclination of the lines would be diminished, since tan ^ 
varies inversely with p, and it could not be increased by employ- 
ing a greater dispersion, as the brightness of the spectrum, which 
would be the same for both telescopes, would hardly bear any 
further reduction. A material advantage would be that with the 
same slit-width a smaller area of the image would be included 
between the jaws, and hence at any part of the slit there would 
be fewer points having different velocities in the line of sight. On 
the whole, it seems to me that the advantage would lie with the 
large telescope. With a reflector, or a photographically cor- 
rected refractor, the photographs could be taken at the Hy line, 



METEORIC CONSTITUTION OF SATURN'S RINGS 427 

where the dispersion is more than twice as great as in the region 
near X 5350, and the only difficulty in that case would be found 
in the yellow color of Saturn. 

I have given a somewhat full account of these observations, 
partly because of the interest inherent in everything that relates 
to the magnificent system of Saturn, and partly because the 
successful application of the spectroscope to the measurement of 
celestial motions depends largely upon details of appliances and 
methods. 



REMARKS ON PROFESSOR E.G. PICKERING'S ARTICLE, 
-COMPARISON OF PHOTOMETRIC MAGNITUDES 
OF THE STARS," IN A. N. 3269.' 

By G. MuLLER and P. Kempf. 

In No. 3269 of the Asiranomische NackrickUn^ Professor E. C. 
Pickering has published a paper in which special consideration is 
given to 38 stars — the stars being those which we had arranged 
in a separate table* in our Photonutric Durchtnusterung^ because 
our value of their brightness differs by at least half a magnitude 
from that of the Cambridge catalogues or the Uranometria nova 
Oxatdensis- The entire method of treatment in this paper, and 
the conclusions which Pickering has arrived at and given in con- 
nection with his discussion, must produce the impression that the 
comparison of these few stars is to be regarded as a sufficient 
basis for estimating the relative accuracy of the different photo- 
metric catalogues in question. This impression is still further 
strengthened by the 4gth Annual Report of the Harvard College 
Observatory^ which has just appeared, and in which the 38 stars 
are again considered. Here it is also directly stated that the 
differences found for these stars serve to indicate the relative 
accuracy of the catalogues referred to. A table is appended to 
the report in which the catalogues are arranged in the ''order of 
excellence,'' and the discrepancies found for the 38 stars are rep- 
resented graphically. The result at which Pickering arrives is that 
the Cambridge catalogue exceeds all of the others in reliability. 

Even if it were not apparent that no conclusion can be drawn 
from such a small number of stars as this as to the degree of 
accuracy of a list containing several thousand stars, we should not 
wish to leave Pickering's article unanswered, especially since we 
cannot by any means approve of the method of treatment which 
he has applied to even the 38 stars in the comparison. 

* Translated from A, N, 3279 at the request of Professor H. C. Vogel. 

• PM, d, AUrophys, Obs, tu Potiddm, g, 500. 

428 



PHOTOMETRIC MAGNITUDES OF THE STARS 429 

Pickering has considered in all five different catalogues, those of 
Oxford and Potsdam, and three others, all of which are Cambridge 
catalogues. These last are the Harvard Photometry, the revision of 
the Boim Durchmusterung {^Harvard Annals, Vol. XXIV) , and a revi- 
sion of the Harvard Photometrywhich has not yet been published. 

It is first of ail to be remarked that 9 of the 38 stars 
cannot be employed in the investigation, since they are included 
in only two of the catalogues. The brightness of three of 
these stars as given in his table is, moreover, admitted by 
Pickering himself to be erroneous, for he assumes that the stars 
No. 1698 and No. 1699 have been transposed in consequence of 
a misprint, and, in the case of No. 1218, for which the difference 
"Potsdam — Pickering" reaches the value of 1.34 magnitude, he 
considers that the observation undoubtedly applies to some other 
star. Moreover, a lack of uniformity is still found in the 29 
remaining stars, inasmuch as some of them occur in only three, oth- 
ers in four, and only nineteen in all five catalogues. Nevertheless 
Pickering takes for each star the mean of such values as are given, 
and regards the residuals obtained by comparison with this mean 
value as a criterion for the accuracy of the different catalogues. 

Such a procedure could be regarded as justified in a certain 
sense, only in case the differences between the individual cata- 
logues could be regarded as purely accidental and not in any 
way systematic. But since a glance at the separate columns of 
Pickering's table shows that there is a marked preponderance of 
one algebraic sign, and since it has moreover been shown in a 
previous comparison ' of the Harvard Pliotometry and the Urano- 
metria twva Oxoniensis that there is a systematic difference 
between these two catalogues, the method of treatment adopted 
by Pickering can hardly be regarded as applicable. 

The comparison in Pickering's article is interesting to us 
because we learn in it that he has lately reobserved the 29 stars 
above referred to, and that his new measures agree so well with 
our own that the mean discrepancy is now only ±0.18 magni- 
tude, while the difference exceeds 0.5 magnitude in only one 

' V.J. S. der asiron. Gaellschaft, az, 257. 



430 G. Mi/LLER AND P. KEMPF 

case.' This we regard as the best kind of proof that the origi- 
nally large number of differences greater than 0.5 is in the great 
majority of cases not to be attributed to our catalogue. An 
explanation is perhaps still required for the stars 245, 18 13, 
2326 and 2395, 'or which the mean values in the difiEerent 
catalogues of Pickering differ by from 0.63 to 0.86 magnitude, 
and further for the stars 122, 185, 653, 3327, 3361. These last 
five stars are found in Vol. XXIV of the Harvard Annals^ not 
only in the principal catalogue (Table I), but in the "Miscel- 
laneous Measurements'' (Table IV), and Pickering has used the 
mean of the values in the two tables. Now since it is expressly 
stated in Vol. XXIV, p. 202, that "When a star was found to be 
common to Tables I and IV, all of the observations were gener- 
ally transferred to Table I," it seems to us that the value in 
Table I only should be used. By taking into account the values 
in Table IV, Pickering obtains further mean values which agree 
with our own to within allowable limits (there remains only one 
difiEerence greater than 0.3), so that for these stars also the 
responsibility for the large discrepancies must be borne by the 
Photometric Revision, 

We believe we have sufficiently proved in the preceding para- 
graphs that Pickering's conclusion, based on the comparison of 
a few stars, that the photometric catalogue of the Harvard 
Observatory is superior to all others, is quite untenable ; we shall, 
however, on our side take this opportunity to make some general 
remarks on the photometric investigations at Cambridge, for 
which the contents of Pickering's article give the occasion. 

In our Photometric Durchmusterung we have already subjected 
the Cambridge catalogues to a stringent criticism and have brought 
forward everything that can be said for or against them. We 
only desire here to again state, and with even stronger emphasis, 
that in our opinion the Cambridge catalogue will always have a 
high value because it gave for the first time a systematic cata- 
logue of the magnitudes of a large number of stars, but that on 
the other hand it has not that degree of accuracy which must be 

' The star No. 653, the striking color of which (G R) at once accounts for the 
large discrepancy. 



PHOTOMETRIC MAGNITUDES OF THE STARS 43 ^ 

Striven for with the instrumental means of today, and which can 
in fact be attained. The reasotis for this opinion rest upon the 
following considerations : 

The most reliable criterion for the excellence of photometric 
observations is afforded by the agreement of the values obtained 
on different evenings for the brightness of the same star. That 
the probable error dz o. i $ of one observation in the Harvard 
Phatonutry is rather large may be seen from the fact that the 
same error for the Potsdam measures is ±. 0.06. The number of 
observations for each star at Cambridge is not so much greater 
than at Potsdam as to reduce the larger uncertainty of the sepa- 
rate measurements to approximate equality in the final results. 
The average probable error of a catalogue brightness is for Cam- 
bridge ±. 0.075 ^^^ ^or Potsdam ±: 0.040. But if the individual 
stars of the Harvard Photometry are considered, it will be found 
that differences of the values obtained on different nights, 
amounting to a whole magnitude or more, occur in the case of 
more than 400 stars (out of 4260), i. r., in about one star in ten. 
For 1 70 stars it even happens that one and sometimes even sev- 
eral measures dS&tx from the mean of the others by more than a 
whole magnitude. These observations are simply rejected. In 
the Photometric Revision the limits of permissible discordance are 
drawn somewhat closer, in consequence of which 5 50 measurements 
in all, differing from the mean of the other observations by at 
least 0.6 magnitude, are rejected. It is evident that so high a per- 
centage of entirely erroneous measurements ought not to occur, 
and that measurements of one and the same star which differ by 
a whole magnitude deserve little confidence. When Pickering 
asserts in his article that the accidental errors of observation are 
of little importance, and then adds, "At Harvard, when an 
observation was discordant, the star was reobserved on so many 
nights that the final value was generally only changed one or two- 
tenths of a magnitude whether the observation was retained or 
rejected," we desire to say, on our part, that such a degree of 
accuracy in photometric measurements seems to us to be entirely 
inadequate. Observations of so small reliability as this are in 



432 G. MULLER AND P. KEMPF 

no way preferable to mere estimates of brightness, in which much 
better results can easily be obtained by a practiced observer. 

As an explanation of the above characteristic discordances 
there only remains, in our opinion, one of the following assump 
tions : Either the stars concerned are variable, or the atmospheric 
conditions on one or the other of the evenings were not above 
suspicion, or finally, mistakes have been repeatedly made in the 
identification of stars. There are no other explanations for such 
enormous differences in photometric measurements, and sincj 
the first two assumptions are very improbable when a consider- 
able number of stars is in question, we are forced to ascribe the 
greater part of the decidedly erroneous observations in the Cam- 
bridge catalogue to the cause last mentioned. We must again 
emphasize the opinion which we have already expressed in our 
Durchmusterung that the Cambridge measurements have been 
made in far too great haste to exclude the possibility of frequent 
erroneous identifications of stars, and this opinion is not shaken 
by the closing remarks in Pickering's article. We admit that in 
specially favorable cases observations of a few stars can be made 
at the rate of one every minute, but we regard it as inconceivable 
that the same rapidity can be maintained as an average for a 
large number of stars without laying the certainty of their 
identification open to question. We cannot regard as advanta* 
geous the measurement by one and the same observer of 63 
stars in 59 minutes {^Harvard Annals^ XIV, series 512), of 201 
stars in 195 minutes (XXIII, series 1015) or of 87 stars in 44 
minutes (XXIII, series 930); and when finally 61 stars are 
observed in 26 minutes (XXIHt series 942), the interval of 26 
seconds available for each star is hardly sufficient for setting the 
photometer with proper care, leaving no time whatever for a cer- 
tain identification of the star. While it is true that rapid proce- 
dure is attended with certain advantages from a quantitative 
standpoint, allowing, for example, a surprisingly quick survey of 
the whole heavens to be made, it is quite unavoidable that when 
observations are made with such haste as this, the quality of the 
results must suffer in a not inconsiderable degree. 



Minor Contributions and Notes. 



THE SHORT WAVE-LENGTHS OF THE SPARK SPECTRUM 
OF ALUMINIUM. 

I HAVE lately determined the short wave-leogths of the aluminium 
spark by means of a short-focus Rowland concave grating. They were 
photographed in the second order together with the overlapping part 
of the first order of the spectrum of iron. Many of the iron lines in 
this part are included in Rowland's new table of standard wave- 
lengths.' From these the wave-lengths of the aluminium lines were 
interpolated. 

The photographs were taken in vacuo and in air of atmospheric 
pressure on plates prepared by Professor Kayser according to Dr. V. 
Schumann's process:' 



Ttfona of metouy nid so* C. 


InTKOO 


Gorans 


1854.09 
1862.20 
»935.a9 
1989.90 


1854.77 
1862.81 
1935.90 
1990.57 


1852.2 
1860.2 

1933.5 
1988. I 



There is another weak aluminium line at about X 1930.4, which I have 
not tried to measure more accurately as it is rather diffuse. Cornu 
makes it X 1928.7. The mean error of my determinations, which are 
the results of six exposures, is less than 0.014. 

C. Rungs. 
Tbchnischb Hochschulb, 
Hannover, Gennany. 

A LARGE ERUPTIVE PROMINENCE. 

On the morning of March 25, 1895, while observing the chromo- 
sphere and prominences, my assistant, Mr. Ferdinand Ellerman, noticed 
at 9^ 50" a rather bright prominence about 2' high at position angle 

* A. and A, is, 1893. 

* V. Schumann : SU». Btr. der Wumr Akad, October, 1893. 
3C0RNU : Jour, di Pkys. 10, 425, 1881. 

433 



434 



MINOR CONTRIBUTIONS AND NOTES 



238°. At lo*" 20"* the Ha line was displaced toward the red, and 
the height of the prominence had visibly increased. Preparations were 
at once made to photograph the prominence with the spectroheliograph 
attached to the 12 -inch telescope. The first exposure was made at lo** 
34*" on plate D3625, and a second was made on the same plate at lo*" 
40"*. At 10^ 46"" a marked displacement toward the blue was noticed 
in the upper part of the prominence, which had now reached a height 
of about 7'. A third exposure was made at 10^ 58*", when the prom- 
inence had probably attained its greatest height. At 1 1** 6"* the upper 
parts had disappeared, and the Ha line could be traced to a distance 
of only about 4' from the limb. It was displaced strongly toward the 
blue at a point near the base of the prominence. 

The three photographs obtained have been reproduced in Plate 
XV, enlarged 3.3 diameters. On this scale the diameter of the Sun 
would be 6.6 inches.. Measurements of the negatives give the follow- 
ing results : 



No. 


Timet 


Height (K line) 


I 
2« 

3 


I0»» 34" 
I0>»40" 
io»»s8» 


300- 
359' 

624' 


135,200 miles 
161.500 " 
280.800 " 



The times given for each photograph were noted when the moving 
Slit of the spectroheliograph had reached a position about 4' from the 
Sun's limb. On account of the slow motion of the slit, and the nature 
of the phenomenon the times thus obtained cannot be relied upon to 
determine the velocity of ascent with accuracy. I hope to have in 
operation soon an apparatus capable of giving complete and exact 
records of such eruptions. 

Mr. Ellerman observed no indication of disturbance On the Sun's 
disk near the position angle of the prominence, and photographs 
taken after the eruption showed no bright faculae at this point on 
the limb. 

George £. Hale. 

' Chicago Mean Time. 

'During this exposure the base of the prominence was accidentally covered by 
the circular metallic disk used to exclude from the spectroheliograph the direct light 
of the photosphere. 



MINOR CONTRIBUTIONS AND NOTEi> 435 

ON A PHOTOGRAPHIC METHOD OF DETERMINING THE 
,. VISIBILITY OF INTERFERENCE FRINGES IN 
SPECTROSCOPIC MEASUREMENTS. 

In a theoretical investigation of the relation between the distribu- 
tion of light in a source as a function of the wave-length, and the 
resulting "visibility curve,"* Professor Michelson has adopted for 
his definition of "visibility" 

"^ /. + /. 

where 7^, I^ are respectively the intensities at the centers of adjoining 
bright and dark interference bands. The curves obtained in the work 
with the interferential refractometer have been platted directly from 
eye-estiinates of the visibility. That such estimates are reliable has 
been clearly shown by Professor Michelson in his important memoir 
"On the Application of Interference Methods to Spectroscopic 
Measurements."* Two quartz lenses, one concave and the other convex, 
and of equal curvatures, were mounted with their axes crossed at right 
angles between two Nicol prisms. The visibility of the concentric 
system of interference rings thus produced depends upon the angles 
between the axis of the quartz and the polarizer and analyzer. If the 
analyzer and polarizer are parallel Professor Michelson has shown that 

/, + -^« I + cos* 2a 

where a is the angle between the axis of the first quartz and the 
principal plane of the polarizer. 

This curve, when platted together with the mean of a number of 
eye-estimates, showed that the error of an estimate was never greater 
than o.i6, while in most cases it was much less than this. "The curves 
show a general tendency to estimate the visibility too high when the 
interference bands are clear, and too low when they are indistinct. 
This tendency may be modified by a number of circumstances — thus, 
it increases with the refrangibility of the light used ; it is greater when 
the field contains a large number of bands than when they are but few ; 
it is greater while the visibility curve is falling than when it is rising; 
it does not seem to be greatly affected by the intensity of the light ; 

^Pkil. ^a^. April, 1891. 

* Smithsonian Conttihtiiant to KntmUdgty No. 842. 



436 MINOR CONTRIBUTIONS AND NOTES 

finally, it varies on different occasions and with different observers. 
Notwithstanding these disturbing causes, the result, after applying the 
correction, will rarely be in error by more than one-tenth of its v&lue, 
and ordinarily the approximation is much closer than this.'" 

There can thus be no doubt that in most applications of interfer- 
ence methods, especially in the laboratory, where the source of light is 
under control, the visibility can be estimated by an experienced 
observer with a sufficient degree of accuracy. Professor Michelson has 
in fact found that no advantage results from matching the fringes with 
a system of artificial fringes in the same field of view, whose visibility 
for any value of the angle a can be taken from a table. The loss of 
time, and the difficulty likely to result from variations in the source of 
light, are not accompanied by a corresponding increase in the 
accuracy of the measures. 

In the infra-red Professor Wadsworth has found that the intensity 
of the fringes can be measured with a bolometer, and a fluorescent eye- 
piece would probably suffice for visual observations of the fringes in the 
lower part of the ultra-violet. In certain cases, however, photographic 
methods seem to offer important advantages. 

In his investigations on elliptic polarization Professor Cornu 
employed photography to record interference fringes obtained with 
ultra-violet light.* A similar method can be very easily applied to 
determine the visibility of the fringes in spectroscopic measurements. 
In my experiments an electric arc taken between horizontal carbon 
poles has served as the source of light. An image of the arc is formed 
on the slit of a concave-grating spectroscope, mounted according to 
Rowland's well-known plan. The second order spectrum of a four- 
inch grating of ten-feet focus, ruled with 14,438 lines to the inch, 
is employed. A small interferential refractometer, similar to that 
described in the papers referred to above, was very kindly loaned to 
me by Professor Michelson. The light from a spectral line enters the 
apparatus through the slit of a small collimator, and finally falls upon 
a photographic plate at the focus of a short telescope. The fringes 
produced with various differences of path have been photographed 
with the light of several lines in the blue part of the spectrum. In 
spite of the great dispersion the exposures need not exceed ten or 
fifteen seconds. The field occupied by the fringes may be widened by 

'/. ^. p. 6. 

' C. R. 108, 917. See also Eder's JahHmch fur Pkatof^raphie^ 1891. 



MINOR CONTRIBUTIONS AND NOTES 437 

opening the first slit of the spectroscope, or, if this is impracticable, 
by moving the collimator lens until the narrow second slit is well out 
of focus. 

The intensity (and hence the visibility) of the fringes can be deter- 
mined from the photographs in a variety of ways. For instance, it 
may be estimated directly, or measured with some form of photometer, 
or determined by comparison with a standard set of photographed 
fringes obtained with quartz lenses and Nicols. 

It is probable that the photographic method will be found to 
be of value in astrophysical work. Observations of the Sun and 
other heavenly bodies, made at the Kenwood Observatory with appa- 
ratus similar to that employed with such marked success by Professor 
Michelson in his interference studies of the spectra of the elements, 
will be described jn a future paper. The fringes have been observed 
without difficulty in the ^a line of solar prominences, and in this case 
their visibility can be sufficiently well estimated. For the H and K 
lines photography will probably give better results. As the time of 
exposure need not be more than a few seconds, the whole process may 
be conducted automatically. It is only necessary to arrange a simple 
series of mechanical or electrical motions to make the exposure, move 
the screw of the refractometer any desired fraction of a turn, and move 
the photographic plate far enough to allow another exposure to be 
made upon it. The observer is thus left free to keep the prominence 
in the proper position on the first slit of the spectroscope. 

It is doubtful whether the fringes could be observed directly in the 
H and K reversals on the solar disk. In any case it would be 
extremely difficult, if not impossible, to estimate their visibility with 
any degree of accuracy. Although the experiment has not yet been 
tried, it seems barely possible that the visibility curves for these 
reversals can be determined by photography. 

In the only attempt yet made to observe the fringes in the chief 
line of the Orion nebula spectrum it was impossible to see them. The 
apparatus was, however, a temporary one, and the silvered mirrors 
were greatly tarnished. Even if the fringes could be seen it would 
hardly be possible to estimate their visibility, on account of their 
faintness. fiut it is probable that, with suitable apparatus, the fringes 
could be photographed, provided the temperature were constant 
during the long exposure required. For such work it might be 
desirable to employ a refractometer with one of its mirrors cut into a 



438 MINOR CONTRIBUTIONS AND NOTES 

number of pieces, each of them adjusted to give a different length of 
path. Thus one or two exposures would suffice to determine the 
visibility curve of the line in question. George E. Hale. 

March x8, 1895. 



NOTE ON THE EXPOSURE REQUIRED IN PHOTOGRAPH- 
ING THE SOLAR CORONA WITHOUT AN ECLIPSE. 

In a recent article "On some Attempts to Photograph the Solar 
Corona without an Eclipse"' I described the Huggins apparatus of the 
Observatory on Mount Etna, and referred to some experiments made 
with it with the object of ascertaining the exposures required for the 
Moon and the solar corona respectively. The corona-like images 
obtained with this apparatus did not appear to me (or to Professor 
Ricc6*) to represent the true corona. Among the reasons advanced 
in support of this conclusion was one that I have since found to be 
inconclusive. I therefore desire in the present note to modify my 
previous statement. 

It seemed to Professor Riccd and myself that the extremely short 
exposure required with the Huggins apparatus must be altogether 
insufficient to cause any impression of the coronal image upon the pho- 
tographic plate, because with the same apparatus the corona could not 
be photographed during an eclipse with many times this exposure. As 
the brightest parts of the corona probably exceed the Moon in bright- 
ness, experiments in photographing the Moon at night sufficed to 
establish the truth of this last assertion. We overlooked the fact, 
however, that the Moon can be photographed in the daytime with an 
exposure much shorter than that required at night. At the Lick 
Observatory Mr. S. W. Burnham photographed the Moon in daylight 
with an aperture of -f^ and an exposure of -phr second.' Using the 
same ratio of focal length to aperture, and a plate of the same make 
and sensitometer number, we have recently been unable to obtain any 
image of the Moon in the first quarter with an exposure less than \ 
second. 

It thus appears that though a feeble light acting upon a photo- 
graphic plate during a given time may produce no developable image, 

^A, and A. October, 1894. 

*See this Journal, January. 1895. 

^LUk Observatory Report on the Total Eelipte of January /, iS9g^ p. 14. 



MINOR CONTRIBUTIONS AND NOTES 439 

yet the same light acting upon the same plate during a much shorter 
time may produce a developable image, provided only that the plate 
be illuminated during the exposure by a second luminous surface super- 
posed upon the first. In order that the image of the first source may 
be visible on the photograph it must exceed the background in den- 
sity by at least one part in sixty (roughly). 

Unfortunately the conclusions reached by Professor Ricc6 and 
myself cannot be sensibly modified by recognition of these facts. The 
independent evidence advanced in the papers mentioned is probably 
sufficient to prove that the corona has not hitherto been photo- 
graphed without an eclipse. The sky near the Sun is so bright that 
the density of the deposit produced by it is not visibly increased by 
the additional deposit due to the comparatively feeble light of the 
corona. 

George £. Hale. 



TERRESTRIAL HELIUM (?). 

The following papers by Professor Ramsay and Mr. Crookes were 
communicated to the Chemical Society at its anniversary meeting. 
Professor Ramsay's paper was as follows : In seeking a clue to com- 
pounds of argon, I was led to repeat experiments of Hillebrand on 
cl^veite, which, as is known, when boiled with weak sulphuric acid, 
gives off a gas hitherto supposed to be nitrogen. This gas proved to 
be almost free from nitrogen ; its spectrum in a Pliicker-s tube showed 
all the prominent argon lines, and, in addition, a brilliant line close 
to, but not coinciding with, the D lines of sodium. There are, more- 
over, a number of other lines, of which one in the green-blue is espe- 
cially prominent. Atmospheric argon shows, besides, three lines in 
the violet which are not to be seen, or, if present, are excessively fee- 
ble, in the spectrum of the gas from cl^veite. This suggests that 
atmospheric argon contains, besides argon, some other gas which has 
as yet not been separated, and which may possibly account for the 
anomalous position of argon in its numerical relations with other ele- 
ments. 

Not having a spectroscope with which accurate measurements can 
be made, I sent a tube of the gas to Mr. Crookes, who has identified 
the yellow line with that of the solar element to which the name 



440 MINOR CONTRIBUTIONS AND NOTES 

" Helium " has been given. He has kindly undertaken to make an 
exhaustive study of its spectrum. 

1 have obtained a considerable quantity of this mixture, and hope 
soon to be able to report concerning its properties. A determination 
of its density promises to be of great interest. 

The spectrum of the gas was next discussed by Mr. Crookes, who 
said : fiy the kindness of Professor Ramsay I have been enabled to 
examine spectroscopically two Plucker tubes filled with some of the 
gas obtained from the rare mineral cl^veite.' The nitrogen had been 
removed by "sparking." On looking at the spectrum, by far the 
most prominent line was seen to be a brilliant yellow one apparently 
occupying the position of the sodium lines. Examination with high 
powers showed, however, that the line remained rigorously single 
when the sodium lines would be widely separated. On throwing 
sodium light into the spectroscope simultaneously with that from the 
new gas, the spectrum of the latter was seen to consist almost entirely 
of a bright yellow line, a little to the more refrangible side of the 
sodium lines, and separated from th^m by a space a little wider than 
twice that separating the two sodium components from one another. 
It appeared as bright and as sharp as D, and D.. Careful measure- 
ments gave its wave-length 587.45 ; the wave-lengths of the sodium 
lines being D„ 589.51, and D„ 588.91. The differences are therefore — 

Wave-leoffths Difliereiices 

D, ... 589.51 



0.60 
1.46 



D. - - - - 588.91 

New line - 587.45 

The spectrum of the gas is, therefore, that of the hypothetical ele- 
ment helium, or D3, the wave-length of which is given by Angstrdm 
as 587.49, and by Cornu as 587.46. 

Besides the helium line, traces of the more prominent lines of argon 
were seen. 

Comparing the visible spectrum of the new gas with the band and 
line spectrum of nitrogen, they are almost identical at the red and 
blue end, but there is a broad space in the green where they differ 
entirely. The helium tube shows lines in the following positions : 

' Cl^veite is a variety of uraninite, chiefly a uranate of uranyle, lead, and the rare 
earths. It contains about 13 per cent, of the rare earths, and about 2.5 per cent, of a 
gas said to be nitrogen. 



MINOR CONTRIBUTIONS AND NOTES 441 

(«) Ds, yellow 587*45 Very strong. Sharp. 

\h) Yellowish green • 568.05 Faint. Sharp. 

(c) Yellowish green 566.41 Very faint. Sharp. 

(d) Green - 516.12 Faint. Sharp. 
{i) Greenish blue • 500.81 Faint. Sharp. 

(/) Blue .... 480.63 Faint. Sharp. 

I have taken photographs of the spectrum given by the helium tube. 
At first glance the ultra-violet part of the spectrum looks like the band 
spectrum of nitrogen, but closer examination shows considerable dif- 
ferences. Some of the lines and bands in the nitrogen spectrum are 
absent in that from the helium tube, whilst there are many fine lines in 
the latter which are absent in nitrogen. Accurate measurements of 
these lines are being taken. 



We reprint from Nature the above account of Professor Ramsay's 
supposed discovery of terrestrial helium, and Mr. Crookes' measures 
of the lines in the spectrum of the new gas. An examination of the 
measures is sufficient to show that the identification of the bright 
yellow line with the D, line cannot yet be considered certain. The 
following measures from Rowland's table of standard wave-lengths 
will be of interest for comparison with the results obtained by 
Crookes. 

Ware -lengths 
D, - 5896.154 



D - - - - 5890.182 

D, . - 5875.9382 



5.972 
14.200 



In a note accompanying Rowland's measure of the D, line it is 
remarked that "this value of the wave-length of D, is the result of 
three series of measurements made with a grating having 20,000 lines to 
the inch, and is accurate to perhaps 0.02." 

Mr. Crookes' value for the difference D,-D, is 6.0 tenth-meters ; 
this is in close agreement with Rowland's value of 5.972. fiut the 
difference D.~New Line, determined by Crookes to be 14.6 tenth- 
meters, agrees very poorly with Rowland's value of 14.200 tenth-meters 
for the difference D.-D,. Whether this discrepancy can be accounted 
for by the comparatively low dispersion probably used by Mr. Crookes 
remains to be seen. 



442 MINOR CONTRIBUTIONS AND NOTES 

A LARGE REFLECTOR FOR THE LICK OBSERVATORY. 

Mr. Edward Crossley, of Halifax, England, proposes to pre- 
sent to the Lick Observatory the three-foot reflecting telescope and 
its dome, which now form part of his private observatory. The grate- 
ful thanks of the Lick Observatory are offered to him for this most 
generous and highly appreciated gift. 

Edward S. Holden. 
Mount Hamilton, 

April 4, 1895- 



Change of Address, — The attention of contributors to The Astro- 
physical Journal is called to the fact that after June 15, 1895, my 
permanent address will be Yerkes Observatory^ Lake Geneva^ Wisconsin. 
All papers for publication and correspondence relating to contribu- 
tions and exchanges as well as all personal communications should be 
sent to this address. George E. Hale. 



Reviews. 



THE SPECTRUM RESEARCHES OF PROFESSOR J. M. EDER 
AND E. VALENTA. 

During the past five years there has been appearing in the Denk- 
schriften der mathem, naturw. CI. der Kais, Acad, der Wissensch. in Wien 
a most important series of papers on spectrum analysis by Professor 
Josef Maria Eder. In 1893 ^^ associated with himself in his work Mr. 
E. Valenta, and all recent papers have been signed by both authors. 

What makes the work of Eder and Valenta especially interesting 
is the fact that their aim has been not to measure the wave-lengths of 
any one substance with extreme accuracy, but rather to investigate 
modifications in the spectra of any substance under varying condi- 
tions and to study the spectra of different compounds of the same 
element, especially in the ultra-violet. Their work has been done with 
extreme care, and the conclusions reached are most important. 

The general method used was to photograph the spectrum under 
investigation and a comparison spectrum of known substances on the 
same plate, the comparison spectrum being along the middle of the 
plate and the other along the two edges. Then, by micrometer 
measurements and a process of interpolation, the wave-lengths could 
be measured. The lenses and prism of the spectroscope were quartz ; 
and the photographic plate was so turned oblique to the axis of the 
" telescope -arm" that it was possible to bring the entire spectrum from 
X2000 to X7600 in perfect focus on one plate. Eye observations were 
of course made in the longer wave-lengths. The comparison spectrum 
generally used was that of the spark discharge between poles made of 
an alloy containing lead, cadmium and zinc. The iron spectrum was 
found to contain too many lines for convenient use. The wave-lengths 
of the various reference lines were at first those given by Hartley and 
Adeney; but later, after the publication of Kayser and Runge's 
measurements, the wave-lengths were reduced to Rowland's scale. 

Eder's first paper,' in 1890, is on the subject of the spectrum of 

burning hydrocarbons ; and he used for this purpose the blue flame of 

the Bunsen burner. He observed of course the well-known carbon 

' Wien. Denksckr. 1890. 

443 



444 REVIEWS 

bands ; but in addition discovered several new ones. He found two 
sets of bands in the same spectrum ; one set, the familiar one which 
appears in the arc-spectrum, has the edges towards the red ; the other 
set (first observed by Eder) has the edges towards the shorter wave- 
lengths and always has a single isolated line near the head of each 
band. He notes that these new bands do not coincide with the "cyan- 
ogen bands/' so-called. In this same paper he mentions some observa- 
tions on the spectrum of water vapor at two different temperatures, that 
of the Bunsen fiame and that of the oxyhydrogen flame. At the lower 
temperature the two bands at X 3064 and X 28 1 1 appear ; and at the higher 
temperature, two other bands, further in the ultra-violet. These bands» 
as is well known, have their edges towards the shorter wave-lengths. . 

In his second paper' Eder gives the results of his investigations of 
the spectrum of the ammonia-oxygen flame, /. e. of ammonia burning 
in an atmosphere of oxygen. He found that it consisted of a series of 
bands whose edges were all turned towards the red. 

In January, 1893, Eduard Valenta's name became associated with 
Eder's in a paper* of great interest on the spark spectrum of carbon 
under various conditions. There was some difficulty in making pure 
carbon conducting; but this was finally overcome by a method 
described by Bunsen. Then, by a most ingenious arrangement, sparks 
were passed between carbon electrodes, dry or wet, in atmospheres of 
hydrogen, carbon -dioxide and air. When the spafk is passed between 
the two poles in hydrogen, the smallest number of lines appear in the 
spectrum, because in other atmospheres like CO., or air, the gas is 
dissociated and various other spectra appear. This fact is of interest 
as indicating the best method of getting spark-spectra of other sub- 
stances, such as sodium ; the carbon poles can be moistened in a 
solution of common salt and then placed in an atmosphere of hydrogen. 
If sparks are now passed, the carbon offers the least possible complica- 
tion. In all the spark-spectra of carbon there were certain character- 
istic lines present which did not belong to the "carbon bands*' or to 
the "cyanogen bands," and these were carefully measured and 
described. This spark-spectrum of carbon is in some respects its 
fundamental one. The carbon bands were generally present also; 
and so were some of the "cyanogen bands" unless the carbon poles 
were entirely free from air. 

In this same paper the spark spectrum of silicon is given. It was 

' IVien, Denksckr, 1890. '/W. 1893. 



REVIEWS 445 

obtained in two ways : one by causing sparks to pass between small 
crystals of silicon imbedded in platinum, and then correcting for the 
platinum lines ; the other by causing sparks to pass between poles of 
carbon moistened with silicon-chloride. There were certain differ- 
ences observed between the spectra obtained in these two ways, espec- 
ially in the broadening of certain lines and in the relative intensity 
of the lines. 

In the fourth paper' the spark-spectrum of boron is given. It was 
obtained by causing sparks to pass between crystals of boron imbedded 
in lead ; and it was found to contain twenty-two lines, lying between 
X3960 and X2000. The two chief lines are at X 2497.7 and X 2496.8, 
and evidently coincide with the only two lines in the arc-spectrum at 
X 2497.82 and X 2496.87 as measured by Rowland and Tatnall. 

In the fifth paper' a description is given of the flame-spectra of 
potassium, sodium, lithium, calcium, strontium, barium and boracic 
acid. The general method used was to feed automatically a fiunsen 
burner with salts of these substances, and to photograph the spectra. 
Sometimes as much as thirty hours' exposure was required. Many 
lines, known in the arc-spectrum, but never before seen in the flame- 
spectrum, were observed and measured. Besides these line-spectra, 
continuous spectia were produced by potassium, sodium and lithium. 
And along with the metallic lines of calcium, strontium and barium, 
band-spectra of their compounds were observed, the most prominent 
of these being oxide bands. Chloride bands were also produced and 
measured. All of these bands as well as those of boracic acid lie 
mainly in the visible spectrum, pointing to a low temperature. 

In the sixth paper,' the ultra-violet absorption of different kinds of 
glass is recorded. Nine kinds of colorless Jena glass, cut in sections 
of i"" and i^ thickness were first investigated. One particular 
kind of light phosphate-crown glass was found to have practically no 
absorption as far as X2500. The absorption of some sixteen kinds of 
colored glass was also studied with special reference to the connection 
between the positions of the absorption bands and the refractive powers. 
Kundt's law that *'the absorption bands move towards the red end of 
the spectrum as the refractive power of the solvent increases" was found 
to hold for many cases. 

The seventh paper ^ contains the comparison of the spectra of potas- 
sium, sodium and cadmium in the flame, the arc and the spark. The 

' min. Dtnkschr, 1893. *I^^' >S93< ^Ihid, 1894. ^ IM, 1894. 



446 REVIEIVS 

flame has a temperature of about looo^ C. ; the temperature of the arc 
cannot be far from 3500® C; and probably the temperature of the 
spark (if it has such a property as a definite temperature) is very high. 
There were observed four lines in the flame-spectrum of sodium and 
twelve in that of potassium. These lines are by far the most prominent 
in the arc-spectra, and are usually easily reversed there. Many more 
lines appear in the arc-spectrum and all of these, with still more, appear 
in the spark-spectrum. The prominent lines in the arc-spectrum con- 
tinue to be the most prominent in the spark-spectrum. The case is 
entirely different with cadmium (as it is also for zinc). Some of the 
chief lines in the arc-spectrum are not present in the spark-spectrum, 
and vice versa. Owing to this fact some mistakes were made by Kayser 
and Runge in identifying certain lines of cadmium selected by Mas- 
cart as standard lines. Mascart used the spark-spectrum and Kayser 
and Runge the arc. This error is pointed out and corrected in this 
paper by Eder and Valenta. Lines which are similar or belong to the 
same series ought to have like changes as the temperature of the spec- 
trum is changed. 

Probably the most interesting paper of all is the last one of July 5, 
1894. This deals with the different spectra of mercury. Observations 
on the arc and spark-spectra and on the ordinary Geissler tube dis- 
charge showed that all three were alike, the most prominent lines in 
one spectrum being also the most prominent in the others. But two 
entirely new spectra were discovered. If mercury vapor is distilling at 
a low pressure through a capillary tube, and if a spark be passed 
through it, spectra are observed which are quite distinct from the 
ordinary one. If there is a large number of Leyden jars in circuit, the 
spectrum consists of an immense number of flne, sharp lines ; but if 
there are no jars in circuit the spectrum is entirely changed ; it becomes 
a series of bands whose edges are towards the red. One spectrum is 
just as complete as the other, neither one being a development of the 
other. The band spectrum corresponds to a trifle lower temperature 
than the new line spectrum ; but it is difficult to see how complexity of 
molecular structure can account for the difference between the two 
spectra in the case of mercury, whose vapor is monatomic. This has, 
of course, a most important bearing on the theory of band and line- 
spectra, and seems to decide deflnitely against some of the present 
ideas concerning them. 

J. S. Ames. 



Recent Publications. 



A LIST ot the titles of recent publications on astrophysical and 
allied subjects will be printed in each number of The Astrophysical 
Journal. In order that these bibliographies may be as complete as 
possible, authors are requested to send copies of their papers to both 
Editors. 

For convenience of reference, the titles are classified in thirteen 
sections. 

I. The Sun. 

Adams, Alex. J. S. On the Connection between Variations of Terres- 
trial Magnetism and Solar Surface Disturbances. Jour. B. A. A. 
5, 260, 1895. 

Brown, Miss E. Third Report of the Section for the Observation of 
the Sun. Mem. B. A. A. 3, Part III, 49-120, 1895. 

LocRYER, J. Norman. Observations of Sun-spot Spectra. Nat. 51, 
448-449. 1895. 

Stratonoff, W. Bestimmung der Rotationsbewegung der Sonne aus 
Fackelpositionen. A. N. 137, 165-167, 189$. 

Tacchini, p. Osservazioni di protuberanze solari fatte al Regio Osser- 
vatorio del CoUegio Romano nel 4° tnmestre del 1894. Mem. 
Spettr. Soc. Ital. 24, 18-20, 1895. 

Tacchini, p. Macchie e facole solari osservate al Regio Osservatorio 
del Collegio Romano nel 4" trimestre del 1894. Mem. Spettr. Soc. 
Ital.a4, 15-17, 1895. 

3. Stars and Stellar Photometry. 

Anderson, Th. D. New Variable Star in Lyra. A. N. 137, 235, 1895. 
Barnard, £. £. Filar Micrometer Measures of Nova Aurigae in 1894. 

A. N. 137, 233-234. 1895. 
Holetschek, J. Beobachtungen des Ver^derlichen W Aquilae. A. N. 

X37i 235-236, 1895. 
Markwick, £. E. On a New Variable Sur in Centaurus. Jour. B. A. 

A. 5, 247-249. 1895- 
Maunder, E. Walter. The Southern Milky Way with the> Sydney 

Star Camera. Knowl. 18, 87, 1895. 
447 



448 RECENT PUBLICATIONS 

OUDEMANS, J. A. C. Ueber die Aenderung der Helligkeit der Fixsterne 

zufolge der eigenen Bewegung in der Richtung der Gesichulinie. A. 

N. 137,169-171. 1895. 
Parkhurst, Henry M. Confirmations of Variability. A. J. No. 340, 

15. 32. 1895. 
Schur, W. Ueber die Beobachtungen von Oertem der grossen Praesep- 

egruppe durch photographische Aufnahmen. A. N. 137, 231, 1895. 
Skinner, Aaron N. New Variable in Cetus, SDM-i^'' 6531. A. J. No. 

342, 15, 48, 1895. 
TissERAND, F. Les variations de lumidre de I'^toile Algol. Bull. 

Mens. Soc. Astr. France, i, 73-77, 1895. 
Waugh, W. R. Third Report of the Section for the Observation of 

Jupiter. Mem. B. A. A. 3, Part IV, 121-150, 1895. 
Yendell, Paul S. On the Variable Stars 7247 RX Cygni and 8116 W 

Cephei. A. J. No. 340, 15, 30-3 ^ "895. 
Yendell, Paul S. Observations of Variable Stars of Short Period 

during the Year 1894. A. J. No. 341, 15, 39, 189$. 
Yendell, Paul S. Observed Maxima and Minima of Long*Period 

Variables, 1894-1895. A. J. No. 341, 15, 35, 1895. 
Yendell, Paul S. Observations of Long- Period Variables. A. J. No. 

340, 15, 28-29, 1895. 

4. Stellar Spectra, Displacements of Lines and Motions in the 

Line of Sight. 
Leduc, a. Note historique sur Tinfiuence du mouvement de la terre sur 
les ph6nomines de la refraction. Jour, de Phys. 4 (3"* S^r.), 106- 
109, 1895. 

5. Planets, Satellites and their Spectra. 

BuCHHOLZ, H. Ueber die Japetusverfinsterung durch Saturn und sein 

Ringsystem vom Jahre 1889. A. N. 137, 241-271, 1895. 
Deslandres, H. Recherches spectrales sur la rotation et les mouve- 

ments des plan^tes. C. R. iio, 417-420, 1895. 
Elger, T. Gwyn. Selenographical Notes. Obs*y 18, 117, 157, 1895. 
Fauth, Ph. Ueber die Verwerthung photographischer Mondaufnahmen . 

A. N. 137, 203-205, 1895. 
Lowell, Percival. Mars — Oases. Pop. Ast. a, 343-349, 1895. 
Lynn, W. T. Cassini and the Principal Division in Saturn's Ring. 

Obs'y 18, 118-120, 1895. 
P01NCAR6, H. Obsenrations au sujet de la communication pr^Mente 

de M. Deslandres. C. R. xso. 420-421, 1895. 



RECENT PUBUCA TIONS 449 

Riccd, A. Eclisse di Luna del 14-1$ Settembre 1894* osservato nel 
Regio Osservatorio di Catania. Mem. Spettr. Soc. Ital. 24, 12-14, 

1895. 
See, T. J. J. Peculiar Illumination of the Moon during the Total 

Eclipse of March 10. A. J. No. 341, 15, 38, 189$. 
Neue Wahmehmungen am Mondkrater Linn^. Sinus, as, 

50-S6. 1895. 

6. Comets, Meteors and their Spectra. 

HoLETSCHER, J. Beobachtungen des £ncke*schen Cometen 189$. A. 

N. 137» 237-238, 1895. 
HussEY, W. J. The Photography of Comets with Notes Concerning 

Comet Rordame. Pop. Ast. a, 353-358, 1895. 
MONCR, W. H. S. The Radiant Points of Meteors. Jour. B. A. A. 5, 

253-256, 1895. 

7. NEBULJB and THEIR SPECTRA. 

Hasselberg, B. Sur les observations spectroscopiques des nebuleuses 
faites k Mt. Hamilton k I'aide du grand r^fracteur de TObservatoire 
de Lick par James £. Keeler. Mem. Soc. Spettr. Ital. 24, i-ii, 

1895. 
Wolf, Max. Notiz tiber die Plejaden-Nebel. A. N. 137, 175. 1895. 

8. Terrestrial Physics. 

Clayton, H. Helm. Eleven-year Sun-spot Weather Period and its 

Multiples. Nat. 51, 436-437, 1895. 
Editor of Nature. The Aurora of March 13. Nat. 519 517-518, 1895. 
Lord Kelvin. The Age of the Earth. Nat. 51, 438-440, 1895. 

9. Experimental and Theoretical Physics. 

Dewar, James. Scientific Uses of Liquid Air. Roy. Inst. Proc. 13 pp., 

1894. 
Rubens, H. Die Ketteler-Helmholtz'sche Dispersionsformel. Wied. 

Ann. 54, 476-485, 1895. 

II. Photography. 

Criswicr, G. S. Development of the Plates of the International Astro- 
graphic Chart. Jour. B. A. A. 5, 245-247, 1895. 



450 RECENT PUBLIC A TIONS 

12. Instruments and Apparatus. 

Grubb, Sir Howard. The Development of the Astronomical Tele- 
scope. Roy. Inst. Proc. i8 pp., 1894. 

Taylor, H. Dennis. An Experiment with a 12^-inch Refractor, 
whereby the Light lost through the Secondary Spectrum is separated 
out and rendered approximately measurable. Mem. R. A. S. 51, 
Part IV., 77-86, 1895. 

Zenger, Ch. V. L*objectif catoptrique et sym^trique. C. R. iso, 609- 
611, 1895. 

13. General Articles, Memoirs and Serial Publications. 

Berthelot. Essais pour faire entrer i*aigon en combinaison chimique. 

C. R. ISO, 581-585, 1895. 
Berthelot. Nouvelles recherches de M. Ramsay sur Targon et sur 

r helium. C. R. iso, 660-661, 1895. 
Berthelot. Remarques sur les spectres de I'argon et de Taurore 

bor^ale. C. R. xso, 662, 1895. 
Ramsay, W. Terrestrial Helium (?). Nat. 51, 512, 1895. 
SiDGREAVES, W. Results of Meteorological, Magnetical and Solar 

Observations. Stonyhurst Coll. Obsy. 84 pp., 1894. 



NOTICE. 

The scope of The Astrophysical Journal includes all investigations of 
radiant energy, whether conducted in the observatory or in the laboratory 
The subjects to which special attention will be given are photographic and 
visual observations of the heavenly bodies (other than those pertaining to 
"astronomy of position"); spectroscopic, photometric, bolometric and radio- 
metric work of all kinds ; descriptions of instruments and apparatus used in 
such investigations ; and theoretical papers bearing on any of these subjects 

In the department of Minor Contributions and Notes subjects may be 
discussed which belong to other closely related fields of investigation. 

It is intended to publish in each number a bibliography of astrophysics, 
in which will be found the titles of recently published astrophysical and 
spectroscopic papers. In order that this list may be as complete as possible, 
and that current work in astrophysics may receive appropriate notice in other 
departments of the Journal, authors are requested to send copies of all 
papers on these and closely allied subjects to both Editors. 

Articles written in any language will be accepted for publication, but 
unless a wish to the contrary is expressed by the author, they will be translated 
into English. If a request is sent with the manuscript twenty-five reprint 
copies of each paper, bound in covers, will be furnished free of charge to the 
author. Additional copies may be obtained at cost price. No reprints can 
be sent unless a request for them is received before the Journal goes to 
press. 

The Editors do not hold themselves responsible for opinions expressed by 
contributors. 

The Astrophysical Journal will be published on the first day of every 
month except July and September. The annual subscription price for the 
United States, Canada and Mexico is $4.00 ; for other countries in the Postal 
Union it is 18 shillings. Correspondence relating to subscriptions and adver- 
tisements should be addressed to The University of Chicago^ University Press 
Division^ Chicago, 111, 

Wm, Wesley &* Son, 28 Essex St., Strand, London, are sole foreign agents, 
and to them all European subscriptions should be addressed. 

All papers for publication and correspondence relating to contributions 
and exchanges should be addressed to George E. Hale, Kenwood Observatory, 
Chicago, III, After June 1$, 1895, Professor Hale's permanent address will 
be Yerkes Observatory, Lake Geneva, Wisconsin, 



451 



INDEX TO VOL. I. 



SUBJECTS. 

PAGB. 

Aluminium, Short Wave-lengths of the Spark Spectrum of. (7. Rungg 433 

T ANDROMEDiE. E, C, Pickering 305 

Aquila and Cygnus, Distribution of Stars and Distance of the Milky 

Way in. C. Easian 216 

Argon, Spectrum of. H. F. Newall 372 

ASTROPHYSICAL JOURNAL. George E. Hale 80 

Atmosphere, Determining the Extent of a Planet's. IV. W. Campbell 85 

Atmospheric Bands in the Spectrum of Mars. IViiliam Huggins - 193 

Auriga, Nova, Recent Changes in the Spectrum of. W. W. Campbell 49 
Beryllium and Boron, Arc-Spectra of. H, A.Rowland and R. R. 

Tatnall 14 

Boron and Beryllium, Arc-Spectra of. H. A. Rowland and R. R. 

Tatnall 14 

Brester*s Views as to the Tranquillity of the Solar Atmosphere. 

Egon von Oppolzer 260 

* Cephei, Spectrum of. A. Bilopolsky 160 

Chicago Academy of Sciences, Meeting of the Sec. of Astr.^ Math. 

and Phys. T. J. J. See 86 

Color of Sirius in Ancient Times. IV. T.Lynn - - - - 35i 

Copper, Arc-Spectrum of. H. Kayser 84 

Corona, Exposure Required in Photographing without an Eclipse. 

George E. Hale 438 

Method of Mapping the. George E. Hale 318 

Some attempts to Photograph the, at Mount Etna Observatory. 

A.Riccd 18 

Cygnus and Aquila, Distribution of Stars and Distance of the Milky 

Way in. C. Easton 216 

Dome, Combination Telescope and. A. E. Douglass - 401 
Eclipse of Jupiter's Fourth Satellite, February 19, 1895. E. C. 

Pickering 309 

Eclipses of Jupiter's Satellites, Photographic Observations of. Wil- 

lard P. Gerrish 146 

Electric Motors for Constant Speed, Design of. F. L. O. Wadsworth 169 

Es.-BiRM. 281, Variability of. T.E.Espin 351 

453 



454 INDEX 

PAGB. 

Germanium, Arc- Spectrum of. H. A. Rowland and R. R, Tatnali - 149 

Helium, Terrestrial 439 

Z Herculis. -.V. C Dunir 285 

Interference Fringes, Photographic Method of Determining Visi- 
bility of. Georgg E. HaU 435 

Jupiter's Fourth Satellite, Eclipse of. E, C. Pickering - - - 309 
Satellites, Photographic Obsefvations of the Eclipses of. Willard 

P. Gerrish - 146 

Lens for Adapting a Visual Telescope to Photographic Observations 

with the Spectroscope. James E, Keeler - - - - loi 

Photographic Correcting, for Visual Telescopes. Jatnes E. KeeUr 350 

Lick Observatory, Large Reflector for the. Edward S, HoUen - 442 
Littrow Spectroscope, Pulfrich's Modification of the. James E, 

Keeler 353 

Longitudes, Martian. Percivai Lowell 393 

Mars, Atmospheric Bands in the Spectrum of. William Huggins • 193 

Cloud-like Spot on the Terminator of. A, E, Douglass - - 127 
Observations of, with the Melbourne Great Telescope. R, L, J, 

Ellery 47 

Spectrum of. Lewis E. Jewell 311 

Martian Longitudes. Percivai Lowell 393 

Meteoric Constitution of Saturn's Rings, Spectroscopic Proof of the. 

James E. Keeler 416 

Milky Way, Distance of. C, Eastan 216 

Photographs of. E, E, Barnard 10 

Modern Spectroscope. X. General Considerations Respecting the 

Design of Astronomical Spectroscopes. F, L O. IVadsworth - 52 
XL Some New Designs of Combined Grating and Prismatic 
Spectroscopes of the Fixed-Arm Type, and a New Form of 

Objective Prism. F, L. O. IVadsworth 232 

XIL The Tulse Hill Ultra- Violet Spectroscope. William Hug- 

gins 359 

Motors for Constant Speed, Design of Electric. F. L, O, IVads- 
worth 169 

Nova AuRiGiE, Recent Changes in the Spectrum of. IV. W, Campbell 49 

Objective Prism, New Form of. F. L. O, IVadsworth - - - 232 

Peculiar Spectra, Stars having. M, Fleming 411 

Photographic Correcting Lens for Visual Telescopes. James E, 

Keeler 350 

Method of Determining the Visibility of Interference Fringes. 

George E. Hale 435 

Observations of Eclipses of Jupiter's Satellites. Willard P. Gerrish 1 46 



INDEX 455 

PAGB. 

Photographic Spectra of Variable Sun. E, C, Pickering - - 27 
Photographing the Solar Corona without an Eclipse, Exposure 

Required. Geargt E. Hale 43^ 

Photographs of the Milky Way. E, E. Barnard - - - - 10 
Photography of the Solar Corona without an Eclipse. A, Riccd - 18 
of the Sun, on the Conditions which Affect the Spectro-. A* A, 

MicheUan 1 

Photometric Magnitudes of the Stars, Comparison of. E, C. 

Pickering I54 

G. Muller and P. Kempf 428 

Pickering's Article " Comparison of Photometric Magnitudes of the 

Stars," Remarks on. G. Muller and P, Kempf - - - 428 
Planet, Displacement of Spectral Lines Caused by the Rotation of a. 

James E, Keeler 352 

Planets, Spectra of the. H. C. Vogel 196, 273 

Planet*s Atmosphere, Determining the Extent of a. W, IV, 

Campbell 8$ 

Prism, Objective, New Form of. F, L. O. IVadsworth • - • 232 
Prominence, Large Eruptive. George E, Hale - - - - 433 
Protuberance Observed December 24, 1894. /• F^nyi - - 212 
Pulfrich's Modification of the Littrow Specroscope. James E, 

Keeler 353 

PuLKOWA Refractor, Spectrographic Performance of the. A* 

BdhpoUky 366 

Ran YARD, Arthur Cowper. George E. Hale 168 

Recent Publications, pp. 93, 189, 270, 3541 447* 

Reviews, pp. 88, 180, 263, 443- 

Rotation of a Planet, Displacement of Spectral Lines Caused by. 

James E, Keeler 352 

Saturn's Rings, Spectroscopic Proof of the Meteoric Constitution of. 

James E, Keeler 416 

Schmidt's Theory of the Sun. E. J. IVilczynski - - - - 112 

James E. Keeler 178 

Silvering Solutions and Silvering. F, Z. O. IVadsworth - - 252 

SiRius in Ancient Times, Color of. IV. T.Lynn - - - - 35i 
Solar Atmosphere, Brester's Views as to the Tranquillity of the. 

Egon von Oppblzer 260 

Observations in 1894. P.Tacchini 210 

Spectrum Wave-lengths. H. A. Rowland I, p. 29; II, p. 131 ; 

III, p. 222 ; IV. p. 295 ; V, p. 377. 
Spectra of Boron and Beryllium. H. A. Rowland and R. R. Tatnall 14 
of the Planets. H C. Vogel 196, 273 



456 INDEX 

Spectra, Stars having Peculiar. M. Fleming 411 

of Variable Stars. E. C, Pickering 27 

Spectral Lines, Displacement of, Caused by the Rotation of a 

Planet. James E. Keeler 352 

Spectro-Bolographic Investigations at the Smithsonian Astro- 
physical Observatory. George E, Hale 162 

Spectrographic Performance of the Thirty-inch Pulkowa Refractor. 

A.BilopoUky .- 366 

Spectro- Photography of the Sun, on the Conditions which Affect the. 

A, A„ Michelson i 

Spectroscope, Adapting a Visual Telescope for Photographic Obser* 

vations with the. James E, Keeler loi 

New Form of. C Pulfrich 335 

Pulf rich's Modification of the Littrow. James E. Keeler - • 353 

Tulse Hill Ultra-Violet. William Muggins ' - - - - 359 

Spectroscopes, Design of Astronomical. James E. Keeler - - 248 

of the Fixed-Arm Type. F. Z. O. Wadsworth - - - - 232 

the Design of Astronomical. F, Z. O, Wadsworth - - > 5 a 
Spectroscopic Proof of the Meteoric Constitution of Saturn's Rings. 

James E, Keeler 416 

Spectrum of Aluminium, Short Wave-lengths of the Spark. C. Rnnge 433 

of Argon. H,F,Newall 372 

of « Cephei. A. Bilopolsky 160 

of Copper. H, Kayser 84 

of Germanium. H, A, Rowland and R, R, Tatnall - - - 149 

of Mars. Lewis E, Jewell 311 

of Mars, Atmospheric Bands in the. William Huggins - • 193 

of Nova Aurigae, Recent Changes in the. W, W, Campbell- - 49 

Plates, Putting Wave-lengths on. Olin H, Bcuquin - - - 166 

Wave-lengths, Tables of Solar. H, A. Rowland I, p. 29; II, 

p. 131 ; III. p. 222 ; IV, p. 295 ; V, p. 377. 

Spot on the Terminator of Mars, Cloud-like. A. E. Douglass • - 127 

Stars, Comparison of Photometric Magnitudes of the. E, C. Pickering 1 54 

G. Miiller and P, Kemp/ 428 

Distribution of in Aquila and Cygnus. C. Easton - - - 216 
Discovery of Variable, from their Photographic Spectra. E. C, 

Pickering 27 

Having Peculiar Spectra. Eleven New Variable Stars. M, Fleming 4 1 1 
Sun, on the Conditions which Affect the Spectro-Photography of the. 

A. A, Mickelson i 

Schmidt's Theory of the. K J, Wilcxyttski - - • - iia 

James E. Keeler 178 



INDEX 457 

PACB. 

Telescope and Dome, Combination. A, E, Douglass - - 401 
Visual, Adapted to Photographic Observations with the Spectro- 
scope. James E, Keeler loi 

Telescopes, Photographic Correcting Lens for Visual. James E, Keeler 350 

TuLSE Hill Ultra-Violet Spectroscope. William Muggins - 359 

Variability of £s.-Birm. 281. T, E, Espin 351 

Variable Star Z Herculis. N. C. DufUr 285 

3416 S Velorum. James E. Keeler 262 

Variable Stars, Discovery of, from their Photographic Spectra. E, 

C. Pickering 27 

Eleven New. M.Fleming 411 

3416 S Velorum. James E. Keeler 262 

Wave-lengths of the Spark Spectrum of Aluminium, Short. C. Runge 433 

on Spectrum Plates, Device for Putting. Olin H, Basquin - 166 
Solar Spectrum. H, A, Rowland I, p. 29; II, p. 131; HI, 
p. 222 ; IV, p. 295 ; V, p. 377. 

WoLSiNGHAM Observatory Circular No. 41. T. E, Espin - - - 87 

For titles of Reviews see table of contents. 



INDEX OF AUTHORS. 



Ames, J. S., Reviews op : 

On the Spectrum of the Electric Discharge in Liquid Oxygen, 

Air, and Nitrogen. Liveing and Dewar - - - . 88 
On Variations Observed in the Spectra of Carbon Electrodes, 
and on the Influence of one Substance on the Spectrum of 

Another. W. N. Hartley 88 

Flame Spectra at High Temperatures. II and III. W. N. 

Hartley 89 

Beitr&ge zur Kenntniss der Linienspectren. J. R. Rydberg - go 
Beitr&ge zur Kenntniss der Linienspectren. H. Kayser und 

C. Runge --qo 

Ueber die Spectra von Zinn, Blei, Arsen, Antimon, Wismuth. 

H. Kayser und C. Runge qi 

The Spectrum Researches of Professor J. M. Eder and E. 

Valenta 443 

Barnard, E. £. Photographs of the Milky Way • - - - 10 
Basquin, Olin H. Device for Putting Wave-lengths on Spectrum 

Plates 166 

BjfeLOPOLSKY, A. The Spectrum of B Cephei 160 

On the Spectrographic Performance of the Thirty-inch Pulkowa 

Refractor 366 

Campbell, W. W. Recent Changes in the Spectrum of Nova 

Aurigae 49 

On Determining the Extent of a Planet's Atmosphere - - 85 
Crew, Henry, Reviews op : 

The Luminosity of Gases. III. A. Smithells ... 266 
Popular Scientific Lectures. Ernst Mach .... 267 

Crookes, William. Terrestrial Helium (?) 439 

Douglass, A. E. A Cloud-like Spot on the Terminator of Mars - 127 

A Combination Telescope and Dome 401 

Dun6r, N. C. On the Periodic Changes of the Variable Star Z 

Herculis 285 

Easton, C. On the Distribution of the Stars and the Disunce of the 

Milky Way in Aquila and Cygnus 216 

Ellery, R. L. J. Observations of Mars Made in May and June, 

1894, with the Melbourne Great Telescope - - - - 47 

458 



INDEX OF A UTHORS 4 59 

PAGB. 

EsPiN, T. E. Wolsingham Observatory Circular No. \i - - - 87 

On the Variability of E8.-Binn. 281 351 

FfeNYi, J. On a Very Large Protuberance Observed December 24, 

1894 212 

Fleming, M. Stars Having Peculiar Spectra. Eleven New Variable 

Stars 411 

Frost, E. B., Reviews of : 

Preliminary Report on the Results Obtained with the Prismatic 
Camera during the Total Eclipse of the Sun, April 16, 1893. 
J. Norman Lockyer. The Total Solar Eclipse of April 16, 
1893. Report on Results Obtained with the Slit Spectro- 
scopes. E. H. Hills 91 

£tude sur le spectre de I'^toile variable I Cephei. A. 

B^lopolsky 263 

Gerrish, Willard p. Photographic Observations of Eclipses of 

Jupiter's Satellites - - 146 

Hale, George E. The Astrophysical Journal 80 

Spectro*bolographic Investigations at the Smithsonian Astro- 
physical Observatory 162 

Arthur Cowper Ranyard 168 

On a New Method of Mapping the Solar Corona without an 

Eclipse 318 

A Large Eruptive Prominence 433 

On a Photographic Method of Determining the Visibility of 

Interference Fringes in Spectroscopic Measurements - . - 435 
Note on the Exposure Required in Photographing the Solar 

Corona without an Eclipse 438 

Terrestrial Helium (?) 439 

Review of : 

Publications of the Lick Observatory, Volume III - - - 180 
HoLDEN, Edward S. A Large Reflector for the Lick Observatory - 442 
HuGGiNS, William. Note on the Atmospheric Bands in the Spec- 
trum of Mars 193 

The Modem Spectroscope. XII. TheTulse Hill Ultra- Violet 

Spectroscope 359 

Jewell, Lewis E. The Spectrum of Mars - - - - - 311 
Kayser, H. Note on the Arc-Spectrum of Copper - • - • 84 
Keeler, James E. On a Lens for Adapting a Visually Corrected 
Refracting Telescope to Photographic Observations with the 

Spectroscope loi 

Schmidt's Theory of the Sun 178 

The Design of Astronomical Spectroscopes .... 248 



46o INDEX OF A UTHORS 

Keeler, James £. The Variable Star 3416 S Velorum - - 262 

Photographic Correcting Lens for Visual Telescopes - - 350 
The Displacement of Spectral Lines Caused by the Rotation of 

a Planet ' 352 

Dr. Pulf rich's Modification of the Littrow Spectroscope - 353 

A Spectroscopic Proof of the Meteoric Constitution of Saturn's 

Rings 416 

Review of : 
The Source and Mode of Solar Energy throu^out the Universe. 
I. W. Heysinger 268 

Kempf, P.» and G. Muller. Remarks on Professor £. C. Pickering's 
Article, "Comparison of Photometric Magnitudes of the Stars," 
in A, N, 3269 428 

Lowell, Percival. On Martian Longitudes 393 

Lynn, W. T. The Color of Sirius in Ancient Times - - - 351 

MiCHELSON, A. A. On the Conditions which Affect the Spectro- 

photography of the Sun i 

MOller, G. and P. Kempf. Remarks on Professor £. C. Pickering's 
Article, " Comparison of Photometric Magnitudes of the Stars," 
in A, N. 3269 428 

Newall, H. F. Note on the Spectrum of Argon . - . . 372 

Oppolzer, Egon von. On Brester's Views as to the Tranquillity of 

the Solar Atmosphere 260 

Pickering, £. C. Discovery of Variable Stars from their Photo- 
graphic Spectra 27 

Comparison of Photometric Magnitudes of the Stars - - 154 

T Andromedae 305 

Eclipse of Jupiter's Fourth Satellite, February 19, 1895 - • 309 

PuLFRiCH, C. On a New Form of Spectroscope .... 335 

Ramsay, William. Terrestrial Helium (?) 439 

Ricc6, A. On Some Attempts to Photograph the Solar Corona with- 
out an Eclipse, Made at the Motmt Etna Observatory - - 18 

Rowland, H. A. Preliminary Table of Solar Spectrum Wave- 
lengths. I, p. 29 ; II, p. 131 ; III, p. 222 ; IV, p. ^5 ; V, p. 377. 

Rowland, H. A. and R. R. Tatnall. The Arc-Spectra of the Ele- 
ments. I. Boron and Beryllium 14 

II. Germanium 149 

RuNGE, C. The Short Wave-lengths of the Spark Spectrum of 

Aluminium 433 

See, T. J. J. Meeting of the Section of Mathematics, Astronomy and 
Physics of the Chicago Academy of Sciences, December 11, 

1894 86 



INDEX OF A UTHORS 46 1 

PACE 

Tacchini, p. Solar Observations Made at the Royal Observatory of 

the Roman College in I 8q4 210 

Tatnall, R. R. and H. A. Rowland. The Arc-Spectra of the Ele- 
ments. I. Boron and Beryllium 14 

II. Germanium - - - 149 

VooEL, H. C. Recent Researches on the Spectra of the Planets. I 196 

n 273 

Wadsworth, F. L. O. The Modem Spectroscope. X. General Con- 
siderations Respecting the Design of Astronomical Spectro- 
scopes 52 

The Design of Electric Motors for Constant Speed - • 169 
The Modem Spectroscope. XI. Some New Designs of Com- 
bined Grating and Prismatic Spectroscopes of the Fixed- 
Arm Type, and a New Form of Objective Prism - - - 232 
Notes on Silvering Solutions and Silvering - - - - 252 
WiLCZYNSKi, E. J. Schmidt's Theory of the Sun - - - - 112 



PERIODICAL 

THIS BOOK IS DUE ON THE LAST DATE 
STAMPED BELOW 



RENEWED BOOKS ARE SUBJECT TO 
lAAA^DIATE RECALL 



Library, University of California, Davb 




/\strop^^ysicc\l 



PHYSICAL 
SCIENCES 
LIBRARY 



Qf5 



LIBRARY 

IWfl'F'iSITY OF CAI.IPORWIA 

rWTc 

212145